CN1205868A - Intravascular ultrasound enhanced image and signal processing - Google Patents

Abstract

A device and method for intravascular ultrasound imaging. An ultrasound signal transmitter and detector is introduced into and may be moved through a bodily lumen. The ultrasound signal transmitter and detector transmits ultrasonic signals and detects reflected ultrasound signals which contain information relating to the bodily lumen. A processor coupled to the ultrasound signal transmitter and detector is programmed to derive a first image or series of images and a second image or series of images from the detected ultrasound signals. The processor is also programmed to compare the second image or series of images to the first image or series of images respectively. The processor may be programmed to stabilize the second image in relation to the first image and to limit drift. The processor may also be programmed to monitor the first and second images for cardiovascular periodicity, image quality, temporal change and vasomotion.

Description

Intravascular ultrasound resembles and strengthens and signal processing

The present invention relates to be used for the apparatus and method of intravascular ultrasound (" IVUS ") image intensification and signal processing, especially relate to IVUS image and signal message are handled, thus the apparatus and method of raising IVUS image quality and utilization rate.

The IVUS image be by near the device being positioned at for example emitter or emitter along or obtain at a branch of ultrasonic energy of the tip of the conduit that inserts blood vessel projection.Ultrasonic beam from device is rotated 360 ° of internal cross section images of formation continuously in blood vessel, just image is formed in the cross section (x-y).According to the structure of concrete device, can in the identical cross-section of device, obtain image, perhaps in the cross section of the cross section that installs slightly forward (just terminal), obtain image.If conduit moves (just along the z-axle) in blood vessel and along conduit, then can form and demonstrate the image of the different sections of blood vessel (a series of continuous cross section).

IVUS can be used for various types of blood vessels, comprises (but being not limited to) tremulous pulse, vein and other vascular surfaces, and can be used for whole body everywhere.

The ultrasonic signal itself that receives (detection) is an analogue signal.This signal is handled last a series of vectors that comprise digital information that form through analog-and digital-method.Each vector is represented blood vessel different angles sector, the ultrasound wave of blood vessel part response just.The quantity of the information word of each vector (axially sampling resolution) and be used to scan the vector number (laterally sampling resolution) of whole vessel cross-sections may be according to the system type that uses and difference.

Digital vector can be placed on earlier have polar coordinate just A (r is in θ) the two-dimensional matrix or square formation.In this polar coordinate square formation, for example, X-axis is corresponding with coordinate r, and Y-axis is corresponding with coordinate θ.Each value in the square formation is to represent the value of the ultrasound wave response intensity of this position (if system is 8, then the scope of this value is 0-255).

Usually polar coordinate array is not transferred to display, because the visual doctor who produces is not easy to understand.(r, θ) information in and is transformed to Cartesian coordinate usually through several the processing stage, for example X and Y coordinates (the A (X, Y)) that is more readily understood of doctor to be stored in polar coordinate array A.Therefore, (X, Y) X-axis of square formation and Y-axis are with the Cartesian coordinate of corresponding vessel cross-sections for A.Information in the flute card square formation may show for the doctor and analyze at last through further handling.Obtain and show image with different speed according to the situation of system.Some systems can obtain and the demonstration image by the video display frequency, for example about per second 30 width of cloth.

The intracoelomic cavity pipe is that the IVUS inspection of a section of blood vessel is generally carried out as follows: the far-end (being downstream) of blood vessel is put into checked part, then conduit is slowly extracted out along body inner coelom (Z axle), the consecutive image of this that obtains section promptly shows continuously like this.In many cases, conduit links to each other with a mechanical stretch device, and this device (is that typical rate is approximately 0.5～1mm/sec) tractive conduit with a constant speed.

In the IVUS imaging system, above-described be used to show body cavity for example the technology of the image of the cross section of blood vessel nowadays be often to use.Yet there is defective in these systems, because they do not comprise any type of for example PS that moves of blood vessel of compensating duct and/or body cavity that is used for.As everyone knows, in the process of IVUS to the body cavity imaging, moving of conduit and/or body cavity always appears.Should mobile may appear on the cross section (X-Y), or move to (Z axle), or moving all appears in two places along vessel axis.The imaging conduit is blood vessel inclination relatively also, and the result becomes image plane to be not orthogonal to Z axle (this moves and will be called distortion).These move is that power and other factors by heartbeat, blood and/or other flow of liquid are produced on body cavity, vasodilation, doctor's reinforcing and patient physiological causes.

In existing IVUS system, when the imaging conduit remains static or when artificial or machinery are slowly extracted conduit out, relatively moving between conduit and the body cavity is profile variation between the consecutive image, just the principal element of the variation that monitors on display and/or film or TV.The variation of this profile is because be much higher than the speed that changes in the actual tissue that causes owing to extracting out owing to move the speed of the variation that causes.

When image be included in the consecutive image between the compensating duct and body cavity relatively move the time, image is stable.Because the IVUS system of Shi Yonging does not have a kind of image stability that carried out now, therefore do not compensate or proofread and correct relatively moving between conduit and the body cavity.As a result, each tissue part is ceaselessly moving and is rotating on display and/or film or TV.This doctor who makes is difficult to the organizational structure of Accurate Analysis on the IVUS dynamic display.And when unsettled IVUS image flows to a kind of Processing Algorithm of handling a series of consecutive images as input, for example 3D rebuilds or during dissimilar filtering, this will cause performance decline and error in judgement or judge inaccurate.

Existing IVUS image device or conduit occur sometimes owing to electricity or mechanical reason cause malfunctioning.This image that will cause showing has discernible and unrecognizable artefact, and has covered real tissue.Usually do not have to determine automatically whether image comprises the method for the artefact of these types, and the artefact of these types will hinder the image of analyzing blood vessel or body cavity.

The change procedure of cardiovascular function is usually expressed as periodically.Detect this one-period and performance so that the relation between the instantaneous position phase of foundation image and the cardiac cycle under its is called cardiac gating.

At present, cardiac gating is by internal signal, and normally ECG signal (electrocardiogram) is realized.Yet, the ECG gate need gather the ECG signal and with IVUS signal staggered (or synchronously).This needs additional hardware/software.

Tissue signature in the IVUS image of blood vessel can be divided into three common classes: body cavity, the just place of blood or other body fluid flow warps; Blood vessel wall; And the perimeter, just EV tissue or structure.In most of IVUS films (image), the feature of blood is fast-changing minute surface pattern.The outside of blood vessel also changes with high instantaneous frequency.At present, the instantaneous performance of pixel and characteristics of organizational structure can't monitor automatically.

The for example endovascular vasodilation of body cavity is defined as for example variation of external caliber of body cavity.This variation can be caused by natural situation or environmental stimuli.Vasodilation can have dynamic element and basic static composition, and dynamic element is the dynamic change of body cavity size just, for example dynamic change of external caliber in cardiac cycle (shrinking and diastole).Basic static composition is for example variation of the basic bore of blood vessel of body cavity.

Vasodilation can be expressed as the quantitative physiological parameter that embodies the body cavity performance, and for example blood vessel changes the ability of bore under certain conditions.The parameter of these types has at present and may provide about the body cavity status information at the importance that still has in the future medical science and diagnosis, for example about the effect of blood vessel with the treatment of being carried out.

IUVS can be used for monitoring vasodilation, because it provides the image of the basic bore of body cavity and its dynamic change.And IUVS can be used for monitoring whether vasodilation is overall (consistent), just, the entire cross section of body cavity with same numerical value along same direction contraction/relaxation.IUVS can be used for also determining whether vasodilation is inconsistent, thereby causes the localized variation of body cavity bore, and just, the performance of the different piece of body cavity cross section is different.

At present, various types of use IUVS supervision vasodilation are manually finished.So very dull, time-consuming, and hinder real time monitoring vasodilation.

The explanation of IUVS image is by the composition of analyzing still image and monitors that their instantaneous performance realizes.Most of IUVS images can be divided into three essential parts.The part of the inside is the flow channel of body cavity, just the material cavity of blood flow warp for example.Be actual vascular around the flow channel, can comprise the vascular of blood vessel and any other health, form (if ill speckle in addition) by multilayer tissue.Its outer hetero-organization of vascular can belong to surrounding tissue, for example the heart among the coronary vasodilator figure.

When dynamic surveillance IUVS film, when just observing, different with the instantaneous performance of the picture element of the outer tissue of vascular with vascular itself corresponding to the material of the vascular of flowing through with the form of film.For example, in most of IUVS films, the feature of the blood of the blood vessel of flowing through is periodically variable minute surface pattern.EV organizing also shows as periodic variation.The instantaneous performance of the pixels of dynamic I UVS image at present can't monitor automatically.

In existing IUVS shows, in system design, take measures sometimes for example to eliminate by a high frequency transient change that many images are averaged.Yet, can not eliminate the appearance of for example bright grey value characteristics of high-amplitude value so sometimes, and it also has interference effect.

The size of the flow channel of body cavity is a crucial Diagnostic parameters.When needs are diagnosed, with manual method for example the doctor determine it.This is to realize by the profile diagram on the border of describing the flow channel in the image on for example on TV static or display device of eclipsed still image.This artificial extraction method is lost time, inaccuracy, and deviation is arranged.

Be useful at present the commercialization imgae processing software of automatic extraction flow channel.Yet these softwares are based on the gray value of forming still image, and do not consider the different instantaneous performance of material, and the blood of the passage of for example flowing through is different with vascular lamina.

In the vascular treatment process, carry out the inspection of IUVS tractive repeatedly in same vessel segment part usually.For example, typical situation is, at first monitor problematic vessel segment part, judge disease (if any), remove the IUVS conduit, consider therapeutic scheme, treat, for example adopt PTCA-" air bag " or Si Tante bracing or strutting arrangement, reexamine the part of treatment then immediately with IUVS, so that the result of appraisal treatment.Judge the result of treatment and the effect of the treatment that thoroughly evaluating is carried out for correct, wish before the relatively treatment and treatment with the image of rear section, these image reflections are along the vessel cross-sections (just corresponding section part) at the same position place of z-axle.In order to compare, must determine that position in the IUVS image film after position and the treatment in the IUVS image film before treatment corresponds to each other.This process is called coupling (aligning), can make before the treatment and the IUVS image after the treatment accurately relatively.

At present, coupling is by monitoring that treatment IUVS extraction film preceding and treatment back vessel segment part carries out, utilize discernible organizational structure labelling, place the image in the film one by one or side by side, they visually being corresponded to each other.This method is extremely inaccurate, and be difficult for accomplishing, because jitterbug, and on display, often rotate everywhere and/or move, this is because visual not process is stable, and because many organizational structure labellings of finding on the extraction of the IUVS before the treatment film are because vascular treatment may destroyed or change.And the orientation of blood vessel and profile may be because the orientation of the relative blood vessel of IUVS conduit and the variation of relative position change, and the IUVS conduit is to cause owing to inserting after self move and treat relative to the variation of the orientation of blood vessel and relative position again.This coupling is manually carried out, and mainly relies on people's visual identity, and is quite time-consuming and inaccurate.

The present invention solves the problem of relevant IUVS imaging system in the market, and adopts prior art that the accurate IUVS image and the image sequence of the tissue of being diagnosed are provided Xiang the doctor, thereby can the more accurate diagnosis and the estimation state of an illness.

The present invention handles IUVS image and signal message, moves the distortion that causes and inaccurate so that remove all kinds by conduit and body cavity.The result of Chu Liing is quality and the utilization rate that strengthens the IUVS image like this.An advantage of the invention is that each IUVS image is stabilized with respect to the image of front, thereby remove the negative effect that any postorder of a plurality of images is handled.If moving in each image all is horizontal, then can full remuneration is described in the image that each obtained move.

The present invention has also guaranteed the volume reconstruction method so that accurately reproduce tissue, because stablize moving of body cavity.The present invention can be used for all kinds of systems' (IUVS or other) that need freeze-frame, and useful really, because moving along body cavity, the probe of this type systematic (for example ultrasonic probe or other probe) causes relatively move (being relatively moving of probe and/or body cavity).

The present invention is used for the detection bodies intracavity by ultrasonic unit ultrasonic waves transmitted signal, the analogue signal that receives is become polar coordinate signal (A (r, θ)), stabilization signal under polar coordinate system, stable polar coordinate are become Cartesian coordinate (A (X, Y)), stabilization signal under Cartesian coordinates is transferred to display to the freeze-frame of Cartesian coordinate form then.Stable image, perhaps under polar coordinate system or under Cartesian coordinates, can be in the demonstration processing that takes a step forward, perhaps they can not be shown.Carry out any point that is converted to Cartesian coordinate and/or stablizes under Cartesian coordinates before or after can stablizing in polar coordinate system.In addition, it is stable or Cartesian coordinate stable to omit polar coordinate, and this is according to displacement in the image that detects and/or the decision of other factors.And, according to skew and/or other factors in the image that detects, can comprise or omit other forms of stable.

For example, can introduce regular mobile stabilization method so that the flute card displacement (X and/or Y direction) in rotatablely moving in the compensating pole coordinate system (angle) or overall vasodilation (in the diastole or the contraction of r direction) and/or the Cartesian coordinates.

Horizontal rule between consecutive image reproduces moves and is called " displacement ", and promptly whole the consistent of image plane that be organized in moves.In order to stablize the IVUS image, the first step of being carried out is " displacement is judged and detected ".Here refer to judge and detect each the displacement between the consecutive image (if any).System can use a processor that a pair of continuous IVUS image is operated, so that determine whether displacement is arranged between this picture group resembles.This processor can use single algorithm or select from many algorithms so that be used for determining displacement.

System utilizes the displacement in the algorithm simulation image, then the image of this displacement and its previous image is compared.Relatively being called between the image approaches operation, is also referred to as coupling in state of the art.System carries out single approximation process to displacement each time.Judge a series of results that approach operation, so that determine and the previous position (direction and size) that does not have the visual immediate displacement image of displacement.Certainly, an images can be with the same manner image follow-up with it relatively.After actual shift was determined, current image became previous image, and next images becomes current image, and repeated the operation of front.

Utilize displacement to judge and detection that system has determined the type of transverse movement, for example rotate, expand, shrink, be shifted (flute card) or the like, and the size and Orientation of displacement.Next step is " shift compensation ".This is meant that system carries out one or a series of operations to continuous IVUS image, so that each images of previous image stability of relative proximity.This is stable to utilize one or more purposes to be to eliminate " shift reverse " of detected displacement.System can comprise a kind of algorithm or select several realizations each " shift reverse " from multiple algorithm.Determine further handle or show before the logic of which type of shift reverse of the actual enforcement of each images is called " logic with shift ".In case the IVUS image stability detection of desired type move, system can be transformed to flute card (or polar coordinate) graphic form so, so that further handle and finally show that for example the doctor can see stable result.Otherwise, if before some other treatment step, stablize, then the image that is produced is invested display with original form or orientation before unstable, user may be can't see stable effect in this case.

Laterally moving between the image may not be regular, but has local characteristics, that is to say that the moving direction and the size of visual different piece is all inequality.In this case, above-described antihunt means or other types method can be carried out on local basis, so that compensate such moving.

The present invention is used for only detecting cardiac cycle by the information that draws from the IVUS image, and does not need external signal, for example ECG.This method comprises approaches operation, approaches operation and also partly is used for stable process.When conduit is static or when controlling IVUS and extracting out, an important function of sense cycle (cardiac gating just) is the image that allows to select to belong to the same phase of continuous cardiac cycle.When image is when being selected from the same phase of successive heartbeat, selects image can stablize various types of cycles according to cardiac gating and move (comprise laterally, along Z-axle and angular movement).For example, can show these IVUS images, and can be by adding and showing that inserting image compensates in the interval between these images.The IVUS image of selecting by this operation can continue transmission so that further handle.

Be used for approaching operation and also can being used to monitor image quality and showing the artefact relevant of cycle detection with the fault of image device and blood processor.

The operation of judging that is used to be shifted can show vasodilation automatically.When vasodilation causes consecutive image because the variation of external caliber and different the time, this operation can be used as antihunt means.When image stability vasodilation and after the variation that brings, this variation has also just compensated so.It perhaps can show the information that changes about bore, because may comprise important physical information.By to polar coordinate just A (r, θ) Biao Shi consecutive image approaches operation and realizes angiectatic supervision.These operations can be decided according to needed information type carrying out between each polar vector (consecutive image) between the whole image or accordingly.Because showing as the consistent of body cavity bore, overall vasodilation changes, so can estimate by the operation that approaches about whole polar coordinate image.Usually, anyly be applicable to that the overall stable operation of polar coordinate representation may be used to estimate overall vasodilation.

In the IVUS imaging procedures, inconsistent vasodilation may appear under certain conditions, and just have only with a certain sector of the corresponding IVUS image of the ad-hoc location of body cavity and move.For example, this may appear at a certain position that tremulous pulse produces speckle, thereby makes tremulous pulse only not produce the regional diastole or the contraction of speckle.When detecting this and move, system can be divided into many sections parts to the ultrasonic signal of expression body cavity cross section, then, uses a certain algorithm respectively each segment signal corresponding to the correspondent section part of contiguous image to be handled.Show the IVUS image that produces then.The stable of this form can use separately, perhaps is used in combination with stabilization technique previously discussed.It in addition, also can show information, because may comprise important physiologic information about the localized variation of external caliber.

The transient behavior of pixel and their architectural feature can be used for showing enhancing and cut apart (body cavity is cut apart) automatically.If monitor image under the condition of freeze-frame, the performance that strengthens demonstration and dividing processing so can be improved.

According to the present invention, can monitor the transient behavior of IVUS image automatically.The information that extracts by such supervision can be used to improve the accuracy of analyzing the IVUS image.By filtering with suppress fast-changing feature, for example, can strengthen people, for example the understanding of the vascular of the image that shows with the film form on still image and the dynamic image as the result of the transient behavior of the outer tissue of the blood of the body cavity of flowing through and vascular.

Automatically cut apart, just distinguish vascular and material, the blood of the vascular of for example flowing through, the instantaneous performance of the architectural feature that the pixel that can comprise according to it forms, by the automatic distinguishing material for example the algorithm of blood realize.The instantaneous performance of extracting from image can be used for several purposes.For example, carry out instantaneous filtering so that strengthen image, and the variation of detection pixel features can be used for automatic distinguishing body cavity and its periphery.

In all IVUS images, be preferably in stablize before or eliminate conduit itself (and image device) in order to monitor from image.Do not eliminate conduit and may weaken stabilization technique and supervision.Eliminating conduit can finish automatically, because its size is known.

The present invention also is used for for example respective frame of two different I VUS extraction series images before treatment and after the treatment of automatic affirmation (i.e. coupling or location) same section part of vascular.For IVUS extraction series image and the 2nd IVUS extraction series image that compares same section part of body cavity, just an IVUS imaging sequence and the 2nd IVUS imaging sequence, for example on TV, the imaging sequence seen on the film or with digital form, the imaging sequence must be synchronously.Can reach this synchronous coupling, comprise that two groups of consecutive images to belonging to two IVUS imaging sequences approach operation.

In an imaging sequence, select one group of consecutive image, be called reference group.This picture group resembles the part of the vascular that all show from two imaging sequences and selects, and it should be the part of not carrying out treating, because the tissue signature of vascular may be owing to treat and change.Another condition of this matching treatment is that two imaging sequences are known, constant and preferably obtain under same extraction speed.

Approach between the image that operates in reference group and second group the image and carry out, second picture group resembles the consecutive image with the similar number that extracts from second imaging sequence.Then, the relative reference group resembles this second picture group and moves a frame, approaches operation repeatedly.Such operation can repeat pre-determined number, and more each time frame move near the result so that determine maximum close degree.Maximal phase short range degree will determine the framing bit between the image of two imaging sequences to move.Can oppositely move first or the second series image, so that can aim at and/or monitor simultaneously corresponding image automatically.

Therefore, can see corresponding image, so that the effect of definite treatment of for example being carried out, perhaps organizational structure is in the variation of a period of time.In addition, dissimilar stable can or carrying out between two sequences in the image of two sequences discussed above, can before this matching operation, among or carry out afterwards.Therefore, two serial images not only can show synchronously, also can show with same orientation and situation by relative another.

Fig. 1 (a) and (b) two-dimensional array or the square formation that is replaced into the digital vector image in polar coordinate system and the Cartesian coordinates be shown respectively;

Fig. 2 is described in the result that the skew between two width of cloth consecutive images in the Cartesian coordinates is judged;

Fig. 3 shows and describes the image that occurs drift phenomenon in polar coordinate system and the Cartesian coordinates;

Fig. 4 has described the effect of image being carried out stable operation (rotation and the skew of flute card);

Fig. 5 has described body cavity local contraction and the diastole of representing with the Cartesian coordinate of the polar coordinate of image and image;

Fig. 6 shows and is divided into a tetrameric frame image, so that handle according to the present invention;

Fig. 7 shows the blood vessel in polar coordinate system and Cartesian coordinates that detects local vascular dilation;

Fig. 8 with the formal description of curve the result of the local vascular dilation that in the crown vascular of reality, monitors;

Fig. 9 describes ECG and cross-correlation coefficient synchronously in the mode of curve chart;

Figure 10 shows the numerical tabular of one group of cross-correlation coefficient (middle row) that belongs to consecutive image (top delegation, sequence number 1 to 10) and the result of inner cross-correlation (below delegation);

Figure 11 shows the curve chart of the cross-correlation coefficient of the artefact in the expression IVUS image;

Figure 12 illustrates a frame IVUS image that is divided into three essential parts: the body cavity of flow of liquid warp; Actual vascular and surrounding tissue;

Figure 13 shows the result of instantaneous filtering;

Figure 14 shows the image that algorithm produced of Automatic Extraction body cavity;

Figure 15 has described in the reference field part (middle string) of time series (left side string), second series image of the first series image and the first serial image image (the right string) with reference field visual corresponding (coupling) partly.

In intravascular ultrasound imaging system (IVUS), ultrasonic unit, for example emitter or transmitter array are launched and are accepted ultrasonic signal, and signal is replaced into the vector that comprises numerical data at last through handling.The ultrasound wave response of each vector representation body cavity different angles sector.The required vector number (laterally sampling rate) of number of data element in each vector (axially sampling rate) and the whole body cavity cross section of scanning depends on employed concrete IVUS system.

Digital vector is put into shown two-dimensional array of Fig. 1 (a) or square formation at first.Usually, this square formation has alleged polar coordinate, promptly coordinate A (Υ, θ).The X-axis of the square formation shown in Fig. 1 (a) is corresponding with the r coordinate, and the Y-axis of square formation is corresponding with the θ coordinate.Each value gray value normally in the square formation, for example if 8, then the gray value scope is 0-255, is illustrated in the intensity of ultrasonic signal of the corresponding position of body cavity.Can be transformed to this polar coordinate square formation and the corresponding flute card square formation of the Cartesian coordinate of vessel cross section then, shown in Fig. 1 (b) with X-axis and Y-axis.Then, can further handle this image and be transferred on the display.Original array and demonstration both can use polar coordinate also can use Cartesian coordinate respectively.Value in the square formation can be the value beyond the gray scale, and for example, they can be chromatic value or other values, and can be less than or more than 8.

In IVUS imaging extraction, body cavity (being called vascular later on) and/or imaging conduit may experience the relative motion of several forms.The motion of these types comprises: the rotation of (1) image plane, the just variation of the θ coordinate of polar coordinate image; (2) the flute card moves, just the variation of the X coordinate of Cartesian coordinate image and Y coordinate; (3) overall vasodilation, feature are the radial contraction and the diastoles of whole blood vessel, and just the consistent of r coordinate of polar coordinate image changes; (4) local vascular dilation, feature are that the size and Orientation of the different piece radial contraction of blood vessel and diastole is different, just the localized variation of the r coordinate of polar coordinate image; (5) local moving, feature is the motion of different tissues, this motion dependence it in image definite position and change; (6) run through planar motion, just perpendicular to or approximately perpendicular to the motion of (at angle) image plane.

Can carry out the stable of successive original picture to above-mentioned preceding 5 kinds of motions, because motion is limited in the transverse plane.The motion of these types can compensate by the current image of each frame of conversion, and obtain stable so that the similarity degree maximum of it and its previous image.Preceding 3 type of motion can be used to approach to operate and stablize, and this operation is view picture image or visual major part relatively.This be because the motion be the overall situation or be well-regulated in essence.The the 4th and the 5th type of motion is by carrying out the stable operation that approaches on the local basis, because the different piece of image shows as different motions.The 6th type of motion can only obtain partially stabilizedly by the approaching operation of carrying out on the local basis, and this is because motion is not confined in the transverse plane.Such motion can utilize the cardiovascular cycle detection to stablize.

Next part will be described the stable method of the overall situation, describe local stable method then.Utilizing the cardiovascular cycle detection to stablize will describe in the part that cardiovascular cycle is discussed.

Stable for reaching the overall situation, uses and severally approach operation and finish the skew judgement.Approach operation and can measure similarity degree between two images.By conversion first image and measure it and the approximation ratio of previous second image, just similarity degree is finished the skew judgement.For example, can be by view picture first image be realized the skew judgement along an axle or the mobile pixel (perhaps a plurality of) of two axles (r in X in the Cartesian coordinates and/or Y-axis or the polar coordinate system and/or θ direction).Just be shifted in case finish conversion, use first image and the second previous image after predefined function compares conversion.Carry out this conversion repeatedly, described first image is moved an additional pixel (perhaps a plurality of) along same and/or other axles at every turn, and use first image and the second previous image after predefined function compares conversion.Obtain after the value of all displacements, use the position of the global maximum that predefined function relatively draws to represent travel direction and size between first image and its second previous image.

For example, Fig. 2 has described the result that the displacement between two width of cloth consecutive images is judged in Cartesian coordinates.Image A is for example previous image in the cross section of vascular of expression one model, in the bottom-right quadrant that is centered close to square formation of model.Image B is the current image of the same model of expression, but this model has moved along direction upwards and left, and is positioned at the left quadrant of the top of square formation.The size and Orientation that move at the vascular center is represented by arrow.Following square formation is C (displacement X, a displacement Y) square formation, is to use certain to approach operation and finishes the square formation that draws after the displacement judgement.

There are many different can being used to finish algorithm or the mathematical function that approaches operation.One of them is a cross-correlation, can use and pay upright leaf (Fourier) conversion.This method is, each width of cloth of current and previous image all comprises for example 256 * 256 pixels, and each images all uses fft algorithm to pay upright leaf (Fourier) conversion.The conjugation of the FFT of current image be multiply by the FFT of previous image, and the gained result uses the IFFT algorithm to pay upright leaf (Fourier) inverse transformation.Use the formula of the cross-correlation of paying upright leaf (Fourier) conversion to be expressed as follows: C=real (ifft2 ((fft2 (A)) ^*Conj (fft2 (B)))) wherein:

The square formation (for example, 256 * 256) of the previous image of A=;

The square formation of the current image of B=(for example, 256 * 256);

The fft2=Two-dimensional FFT;

The contrary FFT of ifft2=two dimension;

The conj=conjugation;

The real part of real=compound expression;

*=element multiply by element; And

C=cross-correlation square formation.

Use the cross-correlation of finishing by pair upright leaf (Fourier) conversion to approach and judge that being actually a kind of is similar to.This is because the mathematical formulae of paying upright leaf (Fourier) conversion interrelates with unlimited or periodic function or square formation, yet square formation (or visual) is limited size in reality, and needs not to be periodic.When using pair upright leaf (Fourier) conversion to carry out cross-correlation calculation, this method hypothesis is periodic on two axles.

Therefore, this formula be one well approximate, and it has reflected the physical location of polar θ axle of image, however it does not reflect the physical location of polar r axle of image, does not reflect the X of Cartesian coordinate of image and the physical location of Y-axis yet.

Utilize FFT to calculate cross-correlation many advantages are arranged.At first, all values of cross-correlation square formation C (displacement X, displacement Y) draw by this basic calculating.And have special-purpose hardware, just pay upright leaf (Fourier) conversion chip or dsp board and be used for effectively finishing the FFT operation.

It is direct cross-correlation that another kind can be used to finish the algorithm that approaches operation, perhaps through normalization or not normalization.This realizes by following operation, promptly each pixel of current displacement image be multiply by in the previous image correspondingly pixel, and all results added, carries out normalization under the situation of normalized crosscorrelation.Displacement each time produce one and, actual displacement is by judging maximum and expression in the displacement.Correlation formula can be expressed as follows: C (displacement X, displacement Y) $=\underset{x,y}{\mathrm{\Σ}}$ B (the x-X that is shifted, y-are shifted Y) * A (x, y)

The formula of normalized crosscorrelation is: C (displacement X, displacement Y)= $\underset{x,y}{\mathrm{\Σ}}$ B (the x-X that is shifted, y-are shifted Y) * A (x, y)/ $\sqrt{\mathrm{\Σ}}$ (B (the x-X that is shifted, y-are shifted Y) * B (the x-X that is shifted, y-are shifted Y)) $\sqrt{\underset{x,y}{\mathrm{\Σ}}}$ (A (x, y) * A (x, y)) is wherein:

The square formation of the previous image of A=;

The square formation of the current image of B=;

*=pixel multiply by corresponding pixel;

All pixel summations in ∑=square formation;

C=contains the result's of all displacements of carrying out to some extent square formation.

Use this direct cross-correlation method can calculate C (displacement X, displacement Y) to all possible displacement X and displacement Y value.For example, if each has 256 * 256 pixels initial square formation A and B, so, must calculate the value of each displacement X and displacement Y, this value all in-127 to+128 scopes, need be carried out 256 * 256=65 altogether, and 536 displacements are calculated, so that calculate might the be shifted value of the X and the Y that is shifted of the institute of C (displacement X, displacement Y).Finish the global maximum of just having determined square formation after these calculate.

Can more effectively finish direct cross-correlation by the number of times that reduces required arithmetical operation.In order to detect the actual shift between the image, needn't calculate each possible displacement X and displacement Y.Find the position of maximum C (displacement X, displacement Y) of might be shifted X and displacement Y just enough.

Can be used to finish the third algorithm that approaches operation and be ask absolute difference and (SAD).This realizes by following operation, promptly each pixel in the image is deducted in another image its corresponding pixel, asks their absolute value, and all results added.Displacement each time produce one and, actual displacement by minimum and expression.Asking absolute difference can be expressed as follows with formula (SAD):

SAD=absolute value (A-B)

This formula also can be expressed as follows:

C (displacement X, displacement Y) $=\underset{x,y}{\mathrm{\Σ}}$ Abs (B (the x-X that is shifted, y-are shifted Y)-A (x, y))

Wherein:

The square formation of the previous image of A=;

The square formation of the current image of B=;

Abs=asks absolute value;

Subtract each other between the-=element;

∑=all difference summations.

Though the precision of each in all these algorithm/formula may be according to the placement of concrete forms of motion that is run into and system and is slightly different, but should be appreciated that the neither one formula can be judged out in advance and can offer the best or the most accurate result.In addition, above-mentioned formula has many kind distortion, and also has other algorithm/formula can be used to finish displacement judgement and alternative aforesaid method/formula.These algorithm/formula also comprise those methods that are used for matching operation commonly known in the art.

Again with reference to figure 2, adopt cross-correlation to finish if approach operation, (C is shifted X so, displacement Y) is called the cross-correlation square formation, and its global maximum (stain by top left side quadrant is represented) will be arranged in apart from the center certain distance of square formation with on this direction (arrow of square formation C), with the center of vascular among the visual B with respect to the position at the center of the vascular among the visual A the same (arrow among the visual B).

Adopt SAD to finish if approach operation, stain is represented global minimum so, be arranged in apart from the center certain distance of square formation with on this direction (arrow of square formation C), with the center of vascular among the visual B with respect to the position at the center of the vascular among the visual A the same (arrow among the visual B).

Rotatablely move and show as along current polar coordinate image θ coordinate moving relative to previous image.Rotatablely moving of current image can detect by asking the maximum that approaches between current polar coordinate image and its previous image.When current image just in time will obtain the maximum value of approaching during the numerical value of reverse mobile actual motion.For example in the image of 256 * 256 pixels, 128 with cross-correlation image (minima of SAD image) in the difference (representing) of peaked θ coordinate with pixel will represent the direction (plus or minus) and the size of rotating.

The angiectatic feature of the overall situation is the entire cross section diastole and the contraction of vascular.Such motion is presented as and manages the axial outside or inwardly motion of r along the pulse in the polar coordinate system image.Vasodilation can be adopted above-mentioned formula or other formula, compensates by the visual reverse diastole of its previous polar coordinate image relatively of current polar coordinate.With angle stable phase ratio, the stable orientation that does not change image of vasodilation is in fact just come the conversion image by expansion or compressed image.

Flute card offset table is shown previous image the moving in X-axis and/or Y-axis of flute card image relative to it.Such motion is eliminated with actual motion side's mobile in the opposite direction flute card image by the edge.Therefore, with the flute card displacement that Cartesian coordinate is represented, can adopt basically and realize stablizing with rotation and the arithmetical operation together of vasodilation stable phase with polar coordinate representation.

The number of times that the required displacement of global extremum (maximum or minima are determined according to approximating function) of using different computational methods can reduce definite C (displacement X, displacement Y) is judged.For example, a kind of method has been utilized motion contrast pole coordinate between the continuous IVUS image and/or quite little usually this fact of whole size of flute card square formation.This meaning only around the square formation center, can calculate C (displacement X, displacement Y) in the quite little part of the X=0 that just is shifted, displacement Y=0.The extreme value of this part is the global extremum of the C (displacement X, Y is shifted) that comprises bigger displacement X, displacement Y value certainly.Can be certainly be exactly the global extremum that the location according to system changes in the smallest region size scope with interior detected extreme value.Rely on square formation C smoothness and near monotonic nature (especially global extremum) can further reduce the number of times that required displacement is judged.Therefore, if the value of a certain position is near local extremum (for example 5 * 5 pixels) in C (displacement X, displacement Y), it may be the global extremum of whole square formation C (displacement X, displacement Y) so.

Reducing the number of times that required displacement judges can realize like this: at first the center (displacement X=0, displacement Y=0) from square formation C begins to receive and seeks, and checks among a small circle, for example around centre in 5 * 5 the pixel coverage.If found local extremum at this among a small circle, it may be the global extremum of whole square formation C (displacement X, displacement Y) so, receives and seeks and can stop.Yet, if on the border of this scope, found local extremum, the X=-2 that for example is shifted, displacement X=+2, displacement Y=-2, or displacement Y=+2 receive repeatedly and seek around this pixel so, up to find greater than (less than) C (displacement X, displacement Y) of values is worth near all.Because in a large amount of images, there is not motion in the image, in these cases, determine that the number of times of the required displacement of global extremum judgement approximately is 5 * 5=25 time, rather than original 65,536 times.

By also can reduce the number of times that required displacement is judged to image sampling.For example, if the image of 256 * 256 sizes is taken a sample every a pixel, they are reduced to 128 * 128 square formations so.In this case, between such square formation, carry out direct cross-correlation or SAD and comprise 128 * 128 operations, rather than original 256 * 256 times, wherein each operation is moved another image relative to an image.The method for simplifying that sampling is estimated as displacement can intersect use with above-mentioned method for simplifying.

Again with reference to figure 2, the result who approaches operation is, the displacement X that draws have on the occasion of, displacement Y has negative value.For freeze-frame B, just in order to compensate along the moving of X and Y direction, logic with shift will be shifted oppositely, and just change their symbol and do not change their size, and to carrying out these shifting functions with the corresponding square formation of visual B.To oppositely move visual B artificially like this, and make the visual relatively A of visual B not be shifted.

The actual numerical value of using in approaching calculating needs not to be the initial value of the square formation that imaging system provides.For example, when being 2,3 or 4 powers or when adopting additive method to handle, can obtain better result to the initial value involution.

Imaging conduit and confining bed occur to stablize the artefact form in all IVUS images.This feature make between the image to approach operating effect not good because it is not the ingredient of vascular tissue.Therefore, before approaching operation, need from each images, to eliminate conduit and relevant thing, compose 0 value just for its pixel.Eliminating these things from image can realize automatically, because the size of conduit is known.

Displacement is judged and carried out can be pattern.Therefore, displacement is judged and execution can be limited to polar coordinate system or Cartesian coordinates respectively, and perhaps displacement is judged and carried out and can carry out polar coordinate system and Cartesian coordinates in turn.At present,, be transformed to Cartesian coordinate then, so be shifted to judge and carry out displacement to be most convenient with identical order because normally at first utilize polar coordinate with the imaging of IVUS system.Yet, can change order and can not produce any negative effect and result.

The displacement judgment processing can be carried out along an axle or two axles.Usually, even when detecting, preferably also carry out two-dimensional shift and judge along the moving an of axle.Displacement is carried out and can be limited to two axles, axle or not along any axle.

Being used to be shifted the zone of judging in the image needn't be identical with the zone of carrying out shifting function.For example, displacement judges and can use the quite little zone in the image to finish, and displacement will be moved whole image according to the displacement that this zone marks and be finished.

Invalid logic with shift is to carry out shifting function like this, i.e. the displacement that every images is carried out (thereby forming stable image) equal and opposite in direction concerning displacement calculating, and direction is opposite.Yet such logic can produce in processing procedure and be defined as the drift result.Drift is such process, displacement accumulative total of promptly finishing and the displacement that produces growth, and whole relatively image of its size or display are very big.Drift can be in the cardiovascular cycle identical part inaccuracy displacement judge or non-horizontal image in the result of motion.When finish the flute card stable after, drift for example can cause that sizable part of image shifts out display.After finishing spin stabilization, drift can cause image to increase rotation on a certain direction.

Fig. 3 is described in the image that occurs drift in polar coordinate and the Cartesian coordinate.What the left side showed is original picture, and the right is to have carried out polar coordinate and the stable same image of Di Ka.The image of noting the right be how counter-clockwise direction wide-angle rotation and relatively the image on the left side move down.In this case, the execution of rotation and the displacement of flute card can not compensate the actual shift of image, even has also increased the inexactness of moving calculating.

Logic with shift must can solve such drift, so that the error that displacement is calculated is minimum.A kind of prevent or the method for restriction drift at least is that the limit that allows shift value is set.This will reduce drift, be cost with the displacement that can not compensate some reality still.Can use additional method to prevent to drift about or make the minimum of drifting about.These can intersect use with the cardiovascular cycle detection method of discussing with the back.

Image description shown in Figure 4 the effect that image is carried out stable operation (rotation and flute card be shifted).The image on the left side is when observed IVUS image coronarius (having removed conduit) on most of zone of conventional display, and the image on the right has been depicted as the image on the later left side of stable operation and how shows.

The image on the left side and the right can be seen some difference in the careful watch circle 4.At first, the image of image relative to the left side on the right has a small amount of clockwise rotation (being the several years).This is the result of spin stabilization.Secondly, the image on the right is relative to visual total the moving to left on the left side.This point can be observed the distance of leaving the pattern edge at every images mesenteron (cavity) and be detected.This is the flute card stable result that is shifted.

Image shown in the watch circle 4 can not be found out the advantage that shows image stability.Yet, monitor that these advantages can be easily seen in the series embodiment of image like this.In not comprising stable display, the position of conduit will always be positioned at the center of display, and tissue signature will move on display and rotate everywhere.On the contrary, on stable display, the position of conduit will be moved everywhere, and tissue signature will keep static substantially on display.User can see that in fact stablizing to increase subsequent processing steps, but actual display will show the processing image that is drawn with image original attitude and orientation (stable).

Fig. 5 has described by the polar coordinate image and has represented along the moving characteristic of r coordinate, just shrinks or expansion along the overall situation of the vascular that moves of polar vector.Fig. 5 also shows with the identical overall situation of flute card image representation and shrinks or diastole.Fig. 5 (a) shows the basic configuration of the vessel cross section of representing with polar coordinate and Cartesian coordinate.Fig. 5 (b) shows the contraction of relative basic configuration.Fig. 5 (c) shows the consistent diastole of relative basic configuration.

Because overall vasodilation is presented as the consistent of external caliber and changes, anyly be suitable for the stable operation of polar coordinate and may be used to determine overall vasodilation, for example overall vasodilation can be by utilizing approaching operation and estimating of whole polar coordinate image.

After the two-dimensional shift judgement was finished, as mentioned above, the maximum value position of square formation C (X displacement, displacement Y) on the θ axle was used for spin stabilization.Make the position of extreme value like this on the r axle, this point can be used for representing overall vasodilation.Therefore, the by-product judged of the overall angiectatic supervision two-dimensional shift that is the polar coordinate image.

Every pair of consecutive image produces an angiectatic value of expression.The size and Orientation that is shifted between the image characterizes the variation of vascular, vasodilation just.Diastole is represented in minus displacement, and positive displacement is represented to shrink.The size of value is represented the size of vasodilation amount.

In some cases, motion or vasodilation may not be unanimity/rules, though this motion is limited in the image plane transverse plane just.In order to determine motion or angiectatic type, can be divided into several parts to image, and each of these parts is carried out the stable judgement in part.Can determine the type that moves by the displacement of checking the appropriate section of shown these parts in previous image.For example, as shown in Figure 6, the image among Fig. 6 (a) can be divided into four parts, shown in Fig. 6 (b).Can be to each judgement that is shifted respectively in four parts.Can discern the type of actual motion to the result of each judgement that relatively is shifted in four parts.Therefore, can change the stable type that is applied according to detected type of sports.

Realize the stable of local motion by the operation of carrying out on the local basis that approaches.The fraction (" searching " zone) of the fraction of previous visual A (" model " zone) and current visual B participates in local stabilized treatment.Sometimes, carrying out the part again after being preferably in that the overall situation is stable and finishing stablizes.

In local stable process, formerly the model zone of image in (A) moved and utilized and approach operation with interior in the seek area scope, compares with the zone of model size in the current image (B).Each pixel in the new freeze-frame that forms (B ') will be given a new value according to the result who searches and approach operation.

Describe local stablizing by following example, wherein the model zone is 1 * 1 pixel area, promptly has only a pixel; The seek area is 3 * 3 pixel areas, and approaching operation is SAD.In the chart below, the value among the A is that 3 pixel and B intermediate value are that 9 pixel is corresponding pixel.3 * 3 pixel values of 9 pixel peripheral region value have also been described in addition are.

Pixel among the pixel B among the A (3 * 3 seek area) B '

(model zone)

1 10 10

3 7 9 50 1

11 7 60

In this example, according to the condition of top description, value is that to use SAD and the value that finds be all pixels comparisons in 3 * 3 seek areas around 9 the pixel for " model " pixel of 3.In the institute of seek area might situation, the upper left corner value of seek area be 1 pixel will reach minimum sad value (| 1-3|=2).Therefore, in the new freeze-frame that forms (B '), be that the corresponding pixel in the position of 3 and 9 pixel will be by assignment 1 with value.

Usually, the size of model and seek area can be approached operation and one changes with employed.The value (described in example) of needing not to be the actual pixels of current visual B compose for the actual value of the pixel of the new freeze-frame that forms (B '), but a certain function of pixel value.Notice that this point is very important, opposite with the overall situation/regular antihunt means promptly as local stable result, variation has taken place in the visual composition just internal relations between the pixel and the relative original picture of they distributions in freeze-frame.Local stable can the image that polar coordinate and Cartesian coordinate are represented being carried out.

Fig. 7 shows the situation of blood vessel in polar coordinate and Cartesian coordinate, has wherein detected local vascular dilation.When detecting local vascular dilation, the some parts of its expression vessel cross-sections is different with the performance of other parts.

Fig. 7 (a) shows the baseline chart of blood vessel before differentially expanding.Fig. 7 (b) shows an example of local vascular dilation.As shown in polar coordinate and the Cartesian coordinate, the performance difference of four different pieces of blood vessel: the bore of two sections parts of blood vessel does not change, and perhaps not have relative their correspondent section componental movements in the image formerly; A section is partly shunk, and perhaps moves upward; And a section partly expands, and perhaps moves downward.

As monitor, overall vasodilation determination methods is not suitable for estimating local vascular dilation, because the manifestation mode of blood vessel is inconsistent.If for example adopt overall vasodilation to judge for the example shown in Fig. 7, then may detect integral body is zero vasodilation, and contraction and diastole just offseted each other.

Therefore, must use local vascular dilation to judge.This can be by just judging respectively that to each θ (or Y) vector vasodilation realizes to each polar vector.Employing approaches operation to the one dimension displacement of corresponding polar vector.For example, if shifting function utilizes cross-correlation, it is that the displacement of what use is made of one dimension realizes that this point has been described in so following operation.C (displacement X, Y) $=\underset{x}{\mathrm{\Σ}}$ B (the x-X that is shifted, y) * A (x, y) wherein: the square formation of the previous image of A=; The square formation of the current image of B=; *=pixel multiply by corresponding pixel; All pixel summations in ∑=polar coordinate square formation; The two-dimentional square formation of C=correlation coefficient.

Can see, each polar vector (θ or Y) is shifted along an axle (X or r).The value that tax is used to be shifted for each vector and judges may not be the actual value of image, but, it for example can be each the pixel assignment in the polar vector near the meansigma methods of lateral pixel it, for instance, just (X Y) will be assigned A (X, Y-1) A (X to A, Y) and A (X, meansigma methods Y+1).(displacement X Y) carries out similar operation to B.Can make that like this cross-correlation method more is not subject to noise jamming.

Formed two-dimentional square formation (C (displacement X, Y)) like this.Each row in the square formation are deposited between current image and the previous visual corresponding polar vector and are approached/be similar to operating result.This operation also can utilize FFT to finish.

Form after the square formation, detect extreme value (maximum in the cross-correlation operation) position in each row.This extreme value place is represented to mate between current polar vector and its previous polar vector.Therefore, the vasodilation in each polar vector can be characterized by moving radially of each special angle sector of blood vessel.

This information can be used to show local vascular dilation, can sue for peace and average some or whole polar vector, so that determine angiectatic meansigma methods, perhaps is used for other purposes.Therefore, by estimating local vascular dilation, can estimate local and overall vasodilation.

In order effectively to use and/or the quantificational expression physiological parameter, angiectatic size must link with the actual bore of blood vessel in some way mutually.Therefore, vasodilation monitors that measure should combine with automatic or manual external caliber measurement usually.

Except the vasodilation of reality, the displacement of flute card also can detect and be vasodilation.This is because flute card displacement generation is along the displacement of r axle and θ axle when with polar coordinate representation.For actual vasodilation and Di Ka displacement are differentiated, the displacement in the flute card image is judged to express not to be had or very little motion.If detect the displacement of flute card, so must be at first stable.Then, can return polar coordinate to the Cartesian coordinate conversion so that the vasodilation evaluation.This can make determines that actual vasodilation is more effective, and more accurate result is provided.

Curve description among Fig. 8 in the crown vascular of human body, monitor local angiectatic result.Carry out twice local vascular dilation in the roughly same section part of vascular and monitor, and obtain comprising 190 successive images of shown in Fig. 8 (a) and 8 (b) (X-axle).The difference of two curves is that the vasodilation evaluation shown in 8 (a) carried out before tremulous pulse treatment, just before the treatment, and the vasodilation shown in 8 (b) is judged and carried out after the tremulous pulse treatment, after just treating.

In each images, judge local vascular dilation in each polar vector, then all detected single shifter-adders and average,, every images characterizes (Y-axle), the active sign of vasodilation just so that being produced single overall vasodilation.

Unit along the Y-axle does not have direct physiological significance, because do not calculate the actual bore of vascular, but the relation of the numerical value among Fig. 8 (a) and 8 (b) is meaningful, because they draw from same vascular.Like this, can draw important information from these data.Notice that treatment vasodilation afterwards is (the maximum diastole from about 40 to about 150) how to increase.Therefore, even represent vasodilation not exclusively quantitatively, also confirmed physiological variation (may be relevant) with treatment.

Only basis is stored in the information monitoring cardiovascular cycle in the IVUS image, thereby cancellation is to the needs of EGC or any other external signal.This means and to set up contact at every images and their each comfortable cardiovascular between the instantaneous position phase in the cycle, and do not need external signal.In case set up this contact, monitored so and can in many application that need cardiac gating, substitute the EGC signal.This supervision can utilize the operation that approaches between the consecutive image to finish.And, same approach operation and can produce about the quality of IVUS image and the information of their performance.

Cardiac cycle shows with the periodicity of some parameter of drawing in the IVUS image to be represented.If monitored the performance of these parameters, so just can determine cardiac cycle.Known that the frame acquisition rate also can be used as instantaneous parameters and determines cardiac cycle.

Approaching property between the IVUS image is the parameter of a clear performance periodic model continuously.This is because motion exists periodically in the image of most of types.Approximating function can form like this, and promptly each value wherein draws from the operation that approaches between a pair of consecutive image.For example, one group of ten images will produce nine values of approaching operation continuously.

Approximating function can approach from the generation of the operation of cross-correlation type, SAD operation or other types the operation of type of functions and draw.When being used for monitoring period, normalized crosscorrelation can produce extraordinary result.

Following formula table illustrates the formula (as the function of N images) of the cross-correlation coefficient that is used to calculate approximating function: correlation function $\left(N\right)=\underset{x,y}{\mathrm{\Σ}}B(x,y)*A(x,y)/\sqrt{(\underset{x,y}{\mathrm{\Σ}}A{(x,y)}^2*\underset{x,y}{\mathrm{\Σ}}B{(x,y)}^2)}$ Wherein: correlation function (N)=each is produced the one dimension function of a value to image; The square formation (N images) of the previous image of A=; The square formation of the current image of B=(N+1 images); *=pixel multiply by corresponding pixel; All pixel summations in ∑=polar coordinate square formation;

Correlation coefficient is the by-product of stabilized treatment, because always will calculate the central value (displacement X=0, displacement Y=0) of normalized crosscorrelation square formation (C (displacement X, displacement Y)).Concerning being used for stable all types of approximating functions, all be like this.Approach the central value (the C X=0 that is shifted, displacement Y=0) of square formation), perhaps cross-correlation or another kind ofly be used for stable operation always can be used to produce approximating function.

Also can from the image of a frame, calculate correlation function, promptly calculate the value of the some C (displacement X, displacement Y) that those displacement X and displacement Y be not equal to the constituting-functions at zero place with respect to the displacement of other frame.Correlation function needn't be made of whole image, but can be only calculated by the part of image, this partial images or corresponding with other partial images or with respect to other parts of image displacement is arranged.

Fig. 9 depicts EGC and cross-correlation coefficient with the method for synchronization with curve.Two curves all resemble relevant with same picture group.Fig. 9 (a) shows the curve of EGC signal, and Fig. 9 (b) shows the curve of the cross-correlation coefficient that is drawn by continuous IVUS image.Trunnion axis shows number of images (190 width of cloth consecutive images altogether).Can monitor, the cross-correlation coefficient function table among Fig. 9 (b) reveals periodic model, and cycle identical (about 6 heartbeats all are shown) of being showed of its cycle and EGC signal among Fig. 9 (a).

The cycle that monitors approximating function may be very complicated, because approximating function does not have typical shape, it may change in time, and it depends on the type of employed approximating function, and it may change or change with the experimenter with the different section parts of blood vessel.

Can use diverse ways to monitor the periodicity of approximating function continuously and automatically.For example, a kind of method is the threshold-type method.This method monitors the value that surpasses so-called threshold value of approximating function.In case detect this value, this method monitors and surpasses this threshold value when again.Cycle is defined as surpassing the time difference of threshold value.Figure 10 shows an example of this method with the form of chart.The figure shows out the one group of cross correlation numerical value (middle row) that belongs to consecutive image (sequence number 1 to 10 shown in the top delegation).For example, if threshold value is set to 0.885, carries out the transition to from visual #2 so and surpass this threshold value the process of visual #3 for the first time.Carry out the transition to from visual #6 and to surpass this threshold value the process of visual #7 for the second time.Therefore, the time period in cycle is for obtaining the time of 7-3=4 images.

It is inner cross-correlation that another kind can be used for from the method for approximating curve extraction cardiac cycle.This method is utilized a section of approximating function, just one group of successive value.For example, in chart shown in Figure 10, described section may comprise preceding four width of cloth consecutive images, just visual #1 to #4.In case after selected one section, it is relevant with self, produce correlation 1.Then, the section cross-correlation of the identical size of taking out in this Duan Yucong approximating function, but moved forward an images.The section that moves forward two images is repeated this operation, or the like.In example illustrated in fig. 10, section { 0.8,0.83,0.89,0.85} will with move forward an images the section { 0.83,0.89,0.85, the 082} cross-correlation, then, section { 0.8,0.83,0.89,0.85} will with the part that moves forward two images { 0.89,0.85,082,0.87} cross-correlation, or the like.The bottom line of chart shows the result of these inner cross-correlation among Figure 10.First value 1 is the result of this part self cross-correlation.Check these cross correlation values so that determine the position of local maximum.In this example, they are positioned at #1 image and #5 image place (their value illustrates with boldface type).The cycle of gained is the position of local maximum and the difference that receipts are sought starting position (for example #1 image).In this example, the cycle is to be the time that the 5-1=4 images is experienced from gathering #1 image to #5 image.In case detect the cycle, utilize around local maximum, for example one of the #5 image section is restarted to receive and seeks.For example in this example, described new section can be the one group of value of approaching that belongs to visual #4 to #7.

Because value of approaching of the image that the characteristic of related calculating itself, the inside cross-correlation operation on sometime need future to collect sometime.Therefore, different with threshold method, approach method needs memory image (in memorizer), and cycle detection is reviewed and carried out.Also can monitor cardiac cycle by approximating curve is transformed to temporal frequency domain through pair upright leaf (Fourier) conversion.In frequency domain, the cycle should be expressed as and corresponding peak value of cycle.Spectrum analysis can detection peak.

Approximating function can provide other information about the IVUS image, and these information can not for example obtain from EGC from external signal, neither obtain from actual image.The characteristic of this function can represent to be used to form the IVUS image of approximating function or some state of partial images.The key character of the approximating function of this state of expression IVUS image is that approximating function exists periodically and " unsmooth " property.Standard I VUS image should show as quite level and smooth and be periodic approximating function, and for example Fig. 9 (b) is shown.

Yet, if there is no " unsmooth " and/or periodically, this may be illustrated in and form some problem in the IVUS image so, just in the image that forms artefact is arranged, and for example, perhaps causes by mechanical breakdown or by electric fault.Following accompanying drawing helps to analyze this point.Figure 11 shows the curve chart of the cross-correlation coefficient that draws from continuous IVUS image.The shape of this figure is identical with cross-correlation curve among Fig. 9 (b), but it is by forming by the different conduit that is used for different experimenters in this example.

In this example clearly, approximating function had not both shown obvious periodic, did not have level and smooth shape yet, but quite coarse or have spike to occur.In this case, the characteristic of approximating curve is inhomogeneous the causing of being rotated by the IVUS emitter, and described IVUS emitter is responsible for launching/gathering the ultrasonic signal that is presented in the image.This class artefact appears at sometimes wherein to be had in the parts of movable machinery IVUS conduit-emitter apparatus.

When thinking approximating function reflection standard imaging condition, can be used for other purposes.This links mutually with maximum value position in each cycle of approximating function.Determine that these maximums are very important for the algorithm of image processing, this algorithm is handled a few width of cloth consecutive images together.Near the image of finding the maximum image tends to each other to have in approaching property of maximum and the little image and moves.And, if select to belong to the image of the same phase place of continuous cardiac cycle, preferably utilize usually in each cycle maximum (approximating function) selection they.

In a kind of display packing, for example, these images are projected onto on the display, and fill visual gap with the interpolation image.By such display packing, can stablize all types of cycle movements.

Also can use the cardiovascular cycle to monitor in the logic with shift stage of stabilized treatment.If avoid drift, the accumulation drift after each section should be very little or be equalled zero cycle time, i.e. all drifts in the one-period section and should equal zero or be approximately equal to zero.This means that utilizing logic with shift and cycle to monitor to combine can eliminate drift phenomenon.

With reference now to accompanying drawing 12,, most IVUS images can be divided into three essential parts.The central area (around the conduit) that is labeled as " body cavity " in Figure 12 is for example body cavity or inner passage (cavity) of the reality of blood flow warp of liquid.Be actual vascular around body cavity, be labeled as " vascular " in Figure 12, this vascular is made up of several layer tissues and speckle (if ill).Is its hetero-organization around the vascular, in Figure 12, is labeled as " outside organization ", just muscle or organ-tissue, the heart in for example crown vascular image.

When dynamic surveillance IVUS image (just with the film form), there are the interior tissue of blood flow and the demonstration of vascular outside organization on every side to show the instantaneous feature different usually with vascular itself.

Automatically the information that the instantaneous performance that monitors pixel in the dynamic I VUS image makes it possible to utilize this method to draw is helped the interpolation of IVUS image.This information can be used for strengthening the IVUS demonstration by the appearance of filtering and the fast-changing feature of inhibition, and wherein fast-changing feature comprises as the liquid of blood and the instantaneous feature of surrounding tissue.This information also can be used for cutting apart automatically, and liquid for example blood and surrounding tissue are distinguished in the instantaneous performance of the characteristics of organizational structure that forms by the composition pixel according to them.Thereby determine the size of body cavity automatically,

Relation between the feature that forms for the corresponding pixel that realizes monitoring automatically instantaneous performance, must judge belonging to consecutive image.It is similar to the employed method of operation of approaching on the localization basis to extract instantaneous performance, as previously mentioned.

Big transient change is characterized by when when an images carries out the transition to another images, the sizable relative gray-value variation of corresponding pixel.These fast transient change can on display, be inhibited, promptly characterize these variations by multiply by original picture with a shielding construction.The transient change of shielding reflection pixel value.The problem that this judgement produces is how to determine whether gray-value variation is owing to the mobile of material or changes that perhaps vascular/conduit is kinetic in corresponding pixel value.Overcome or reduced this problem at least by freeze-frame being carried out this judgement.

Introduce following definition:

B=current (stable or unstable) image;

A=previous (stable or unstable) image;

C=follow-up (stable or unstable) image;

Abs=asks absolute value;

The square formation of using can be Cartesian coordinate or polar form.

The operation that produces square formation D1 below should be defined as follows: D1 is a square formation, wherein each coordinate be the pixel value of X, Y be it be respectively the peripheral region among a small circle that from visual A and B, extracts, for example the absolute value difference of 9 elements (X-2:X+2,3 * 3 squares of Y-2:Y+2-a) and.

For example, following description shows respective element (boldface type) and their near zone among square formation A and the B.

A B D1

1 4 51 3 6 8

6 7 15 3 4 70 190

3 5 83 2 1 6

Position and value are the value below the corresponding pixel of pixel of 4 (in B) and 7 (in A) will be endowed among the square formation D1:

abs(1-3)+abs(4-6)+abs(51-8)+abs(6-3)+abs(7-4)+abs(15-70)+abs(3-2)+abs(5-1)+abs(83-6)=190

D2 defines in a similar manner, but is aimed at square formation B and C.

In fact, square formation D1 and D2 are the difference square formations, are promptly averaged in 3 * 3 zones, so that eliminate local vibration or noise.Big gray-value variation will be expressed as sizable value among square formation D1 and the D2 respectively between image A and the B or between visual B and the C.

Next step forms a new square formation Dmax, and wherein each pixel is the maximum of the corresponding pixel among square formation D1 and the D2:

Dmax=max (D1, D2) wherein: max (D1, D2)=each pixel among the Dmax is the maximum of two corresponding pixel among square formation D1 and the D2.

Therefore, single square formation Dmax has strengthened big pixel variations between square formation A, B and the C especially.Then, by normalization just each pixel of Dmax form shielding square formation (MD) divided by the maximum of Dmax from Dmax.Therefore, the scope of the pixel value of shielding square formation MD is from 0 to 1.

The task of shielding is to multiply by current visual B in the following manner, forms a new square formation or image, is defined as BOUT:

BOUT=(1-MD ⁿ) * B wherein: the original current image of B=; The image that BOUT=is new; ⁿEach pixel among the=square formation MD is from taking power n time, and n is a numerical value normally, and for example value is 2-10.1-MD ⁿ=one square formation, wherein the value of each pixel is 1 to deduct the value of corresponding pixel among the square formation MD.

By carrying out (1-MD ⁿ) calculate, among the MD the slow variation characteristic of reflection little value at 1-MD ⁿIn become big value.And the chance of having only slow variation characteristic will have big value has increased, because big MD value has strengthened (forming square formation MD with the maximum between square formation D1 and the D2) in front.

Shielding (1-MD ⁿ) multiply by current visual B, form new visual BOUT, the probability of the pixel appearance that wherein slowly changes has increased, and fast-changing pixel value has reduced.Numerical value n determines the intensity to the quick variation characteristic inhibition that will see on display.

Figure 13 shows the result of instantaneous filtering.The image on the left side is the original I VUS image (square formation B just) of crown vascular, as what will see on current display.The image on the right is handled through above-mentioned treatment step, just instantaneous filtering (square formation BOUT).Notice that on the right blood and surrounding tissue have been filtered (being suppressed) in the image, thereby the border of body cavity and vascular is easier to identification.

Automatically cut apart liquid for example blood and outside organization and blood vessel wall difference come, this is based on the difference between the instantaneous performance of tissue signature.Under the situation of instantaneous filtering, this method draws by the relation between the corresponding pixel of several successive image.If cause pixel value to change owing to move in the image, the function of algorithm will descend so.Before cutting apart automatically, stablize and to overcome or reduce this problem at least.

Under the situation of instantaneous filtering, introduce following definition:

B=current (stable or unstable) image;

A=previous (stable or unstable) image;

C=follow-up (stable or unstable) image;

Square formation can be Cartesian coordinate or polar form.

Tissue signature can limit as follows: default is " b ", " c ", " d " and " e " for four pixels the most contiguous of the pixel of " a ", and the classification of " a " will be according to its relation with " b ", " c ", " d " and " e " so.This can represent with following description:

b

c?a?d

E may form following classification so: in vertical direction: if a＞b and a＞e so " a " be divided into and belong to the classification I; If a＞b and a＜e " a " so are divided into and belong to the classification II; If a＜b and a＜e " a " so are divided into and belong to the classification III; If a＜b and a＞e " a " so are divided into and belong to the classification IV; If a=b or a=e " a " so are divided into and belong to the classification V; In the horizontal direction: if a＞c and a＞d so " a " be divided into and belong to the classification I; If a＞c and a＜d " a " so are divided into and belong to the classification II; If a＜c and a＜d " a " so are divided into and belong to the classification III; If a＜c and a＞d " a " so are divided into and belong to the classification IV; If a=c or a=d " a " so are divided into and belong to the classification V.

Verticals categories and horizontal classification combine and form a new classification then.Therefore, pixel " a " may belong to 5 * 5=25 possibility classification now.The tissue signature that this means " a " is characterized by one of those (25) individual classifications under it.

For example, in the near zone below:

7

10 10 14

3

Pixel " a "=10 is divided in the classification of classification V (because of 10=10) of the classification I that comprises vertical direction (because 10＞7 and 10＞3) and horizontal direction.Yet, if pixel " a " is to be positioned at following peripheral region:

7

11 10 14

3 it will be divided the classification that belongs to different because its horizontal classification is classification III (10＜11 and 10＜14) now.

By determining each pixel and near its relation of pixel on every side, drawn the characteristic of organizational structure, this characteristic is divided into 25 kinds to each pixel may classifications.The number of classification can change (increase or reduce), for example, divides class condition by changing, and the number of pixel for example can use 8 rather than 4 around the similar use.

The ultimate principle of utilizing the variation zone separatory style of organizational structure such as blood and vascular is the variation that monitors classification under the corresponding pixel of consecutive image.In order to accomplish this point, determine the affiliated classification of each pixel among square formation A, B and the C respectively.Then, whether the classification of checking corresponding pixel has respectively changed.If changed, think that then this pixel is a liquid, for example blood or surrounding tissue pixel.If do not change, so just think that this pixel is the vascular pixel.

Following Example is represented the pixel of continuous square formation A, B and three corresponding pixel of C (value is 8,12 and 14) and their peripheral regions.

A B C

5 9 1 9?8?11 19?12?13 21?14?17

23 100 20

In this example, what value was the classification of 12 pixel (in B) with in A and C is identical, so it is divided into the pixel of the very big blood vessel wall pixel of probability.Yet, if pixel location (20 among the C becomes 13) as follows:

A B C

5 9 1

9?8?11 19?12?13 21?14?17

23 100 20 the classification of the pixel 12 among the pixel 8 among the A and the B is identical so, but the classification of 14 among the C is different with the example of front.Therefore, the pixel 12 among the B is divided into the very big liquid of probability (in the body cavity), for example pixel of blood or outside organization's pixel.

Present described sorting technique is observed the organizational structure relevant with each pixel zonule on every side or the variation of figure.In case as mentioned above, determine this variation, composed a binary value then can for each pixel.For example, if determine that it is the vascular pixel, then composing and give numerical value 0, is blood pixel or the pixel that belongs to vascular outside organization if determine it, then composes and gives numerical value 1.Binary image is as the input value of the processing of distinguishing body cavity, and original pixel value no longer plays a role in dividing processing.

Use binary image difference body cavity to be based on two hypothesis, these two hypothesis are effective in the IVUS image of handling in the above described manner.First hypothesis is that the zone or the binary value that is characterized as in the zone of vascular that comprise blood in the image are the density of 1 pixel higher (perhaps value is that the density of 0 pixel is low).Need a boundary density, because pixel is always arranged by misclassification.Second hypothesis is, angle from tissue signature, can the value of seeing be that 1 the pixel high density zone (body cavity) that links to each other is centered around around the conduit, and the quilt value is centered on by 1 pixel low-density continuous regional (vascular), then, centered on by the high zone of 1 PEL density (outside of vascular) by the value that links to each other again.The reason of this hypothesis is that typical organizational structure is to be derived by blood vessel.

These two hypothesis have constituted the basis of following Processing Algorithm, and this algorithm extracts the actual area relevant with body cavity from binary image.This algorithm can use known image processing technique, for example to density feature setting threshold in the localized areas (difference blood/outside organization and vascular) and form and action, for example expansion or link to each other with interior tissue forms the relevant range of representing the actual body cavity of finding in the blood vessel wall scope.

Figure 14 shows the image of Automatic Extraction body cavity algorithm gained.This image is original I VUS image (for example, visual as described above B), and the border of body cavity and bright line overlapping (by described algorithm).The algorithm that extracts the body cavity border is based on using three width of cloth consecutive images to monitor the variation of above-mentioned characteristics of organizational structure.

The example of above-described instantaneous filtering and cutting apart automatically comprises except when also gather the additional images of two width of cloth (for example above-described visual A and C) outside the preceding image (for example above-described visual B), yet these two kinds of methods can be improved to utilizes less (just having only the additional image of a width of cloth) or more additional image.

If combine with the supervision of cardiovascular cycle, the performance of above-described two kinds of methods will strengthen greatly.This is particularly useful for consecutive image, and the cardiovascular cycle in the wherein said consecutive image monitors and produces the value of approaching in the high image.These images do not have motion in the image usually.Therefore, the consecutive image that approaches in having maximum image feedback inputs to instantaneous filtering or will produce the most reliable result when cutting apart automatically.

In the process of using catheterization treatment vascular, common operation is to carry out IVUS repeatedly in same section part of vascular to extract inspection out.For example, typical situation is at first to monitor problematic section part, judges the state of an illness (if ill), remove the IVUS conduit, consider the treatment measure, treat, and then use IVUS to check the section part (at one time) of treatment immediately, so that estimate the result of treatment.

For the correct result who estimates this treatment, the corresponding section part in before the treatment at the same position place of length of tube direction more along the pulse and treatment back.Following method is used for coupling, just discerns (aligning) correspondent section part automatically.

In order to finish the coupling of correspondent section part, belonging to first first group of consecutive image extracting the series image out just between the image of reference section and belong to second and extract out between the image of second group of consecutive image of series image and approach/similar operations.When a certain canonical function reaches maximum, the reference section of the first series image and the second series image on its correspondent section partly reach coupling.

The reference field part is selected from one of described two series images.The reference field part can be one group of visual consecutive image of series of for example representing the IVUS image of several seconds kinds.Importantly select the reference field part in a certain position of vascular, this position all exists in two serial images, and does not change because of any processing, determines that just reference field partly is positioned at the near-end or the top of treatment section part.

As an example, the table among Figure 15 will help to understand the matching process of correspondent section part.

Left side string is represented the time series of the first series image, should comprise 20 width of cloth consecutive images by the series image in this case.Middle string is represented the reference field part selected from the second series image to comprise 10 width of cloth consecutive images.The right string has been listed 10 width of cloth consecutive images (#5-#14) of the first series image, it in fact with the image (#1-#10) corresponding (coupling) of the reference field part of second series image.The purpose of matching treatment is exactly actual this corresponding relation that finds.

In case selected the reference field part, then move it along another series image, a mobile at every turn width of cloth (or several) image, between the corresponding image of each section part, carry out series of stable and approach operation.Moving direction is determined according to the relative position of reference field part in the time series of two series images.Yet, if do not know to move this point usually along both direction.For example, wherein: r=reference field part; And the f=first series image, the first serial operation will be carried out between the visual right image below comprising: r#1-f#1, r#2-f#2, r#3-f#3 ..., r#10-f#10.

The second series operation will be carried out between the visual right image below comprising: r#1-f#2, and r#2-f#3, r#3-f#4 ..., r#10-f#11.

The 3rd serial operation will be carried out between the visual right image below comprising: r#1-f#3, and r#2-f#4, f#3-f#5 ..., r#10-f#12, or the like.As seeing in this example, only there is an images only to be shifted along a direction at every turn.

For example, the operation below can between every pair of image, carrying out.At first, its corresponding image is stablized at rotation and the motion of flute card in an images relative first serial image of reference section.Then, between every pair of image, approach operation.For example, this operation can be normalized cross-correlation (top discussed in conjunction with cycle detection).Each such operation produces the value of approaching, and produces normalization coefficient when for example using normalized crosscorrelation.A series of such operations will produce many such values.In the example shown in the table of Figure 15, the reference field part that at every turn is shifted will produce 10 new cross-correlation coefficients.

Then, the approximating function of the value of the approaching substitution type that can produce sequence of operations is for example in the mean function.Example above using, cross-correlation coefficient is sued for peace then divided by visual logarithm, and just 10.Thereby each group operation produces a single value, and on behalf of the reference field part and the first series image, the average value of approaching just, this value can go up close degree between its interim corresponding image.Therefore, the result of the operation of first series is a single value, and the result of second series operation will be another value, or the like.

We can expect to occur maximum average value and approach, and it is very alike, the result who operates between the just corresponding or matching section part.

In the above example, these sections part should be mated in the 5th serial operating process, and described the 5th series operates in and will carry out between the visual right image below comprising: r#1-f#5, r#2-f#6, r#3-f#7 ..., r#10-f#14.

Therefore, maximum average value approaches and embody corresponding section part, because in fact every pair of image is corresponding image, just they represent same tissue.Yet canonical function can be not according to this algorithm.For example, it can consider the form of approximating function, derives from the position of many shifted segments parts, rather than only uses the value of its maximum.

In case identify the correspondent section part, then make the whole first and second series images synchronized with each other.This is the result who is shifted by the suitable frame that the visual matching treatment of relative another series of serial image draws.Thereby when two series images of supervision arranged side by side, the section part before the treatment will occur simultaneously with the section part after the treatment.

Except make correspondent section part synchronously, aforesaid operations also makes the correspondent section part stable toward each other.This has further strengthened the ability that tissue changes of understanding.Thereby even its position and orientation have changed probably when conduit inserts in the vascular again, still, the image in the serial image in before the treatment and treatment back will be stablized toward each other.

The number of images that is used for the reference field part can change.Number of images used in matching treatment is many more, its stable more influence that is not subject to local error more.Yet the cost of doing like this is that the calculating of each matching process needs more computation time, because visual logarithm has increased.

It is known extracting the stable importance of speed maintenance when obtaining extraction series image out.Preferably extracting speed out is the same to twice collection image.

The present invention can carry out many distortion.Above-described different characteristic can independently and each other make up to contact.These features also can adopt different combinations.

Claims (144)

1. an intravascular ultrasound image device comprises

Place endoceliac ultrasonic signal generator and detector; And

With the processor that ultrasonic signal generator links to each other with detector, this processor is program control to be finished:

A. by deriving first image in the ultrasonic signal that detects,

B. by deriving second image in the ultrasonic signal that detects,

C. compare first image and second image, and

D. handle first image and second image.

2. according to the intravascular ultrasound image device of claim 1, wherein relatively first image comprises the relative first image judgement, second image with second image.

3. according to the intravascular ultrasound image device of claim 1, wherein program control processor of deriving image comprises one of processing and digitized at least.

4. according to the intravascular ultrasound image device of claim 1, also comprise the display that links to each other with processor.

5. according to the intravascular ultrasound image device of claim 1, wherein derive image and comprise the formation two-dimensional array.

6. according to the intravascular ultrasound image device of claim 5, wherein said two-dimensional array is to form in one of polar coordinate system and Cartesian coordinates at least.

7. according to the intravascular ultrasound image device of claim 5, wherein said two-dimensional array is to form in polar coordinate system and Cartesian coordinates.

8. according to the intravascular ultrasound image device of claim 5, wherein said two-dimensional array has a plurality of elements, and each representative in a plurality of elements is from the predetermined detected ultrasonic signal in locus.

9. according to the intravascular ultrasound image device of claim 2, wherein relative first image judges that second image comprises the judgement displacement.

10. according to the intravascular ultrasound image device of claim 2, wherein relative first image is judged that second image comprises and is at least once approached operation.

11. according to the intravascular ultrasound image device of claim 10, the operation that at least once approaches wherein comprises one of cross-correlation, normalized crosscorrelation and SAD at least.

12. according to the intravascular ultrasound image device of claim 11, cross-correlation wherein comprises one of direct cross-correlation and Fourier transformation at least.

13. according to the intravascular ultrasound image device of claim 2, wherein relative first image is judged that second image one of utilizes in polar coordinate system and the Cartesian coordinates at least and is finished.

14. according to the intravascular ultrasound image device of claim 2, the wherein relative first image judgement, second image is realized at least in one direction.

15. according to the intravascular ultrasound image device of claim 1, the wherein further program control detection of processor flute card displacement at least, rotation are moved and one of vasodilation.

16. according to the intravascular ultrasound image device of claim 15, wherein at least displacement of flute card and rotation one of to move be regular.

17. according to the intravascular ultrasound image device of claim 15, wherein at least flute card displacement and rotation one of to move be partial.

18. according to the intravascular ultrasound image device of claim 15, wherein vasodilation is overall.

19. according to the intravascular ultrasound image device of claim 15, wherein vasodilation is partial.

20. according to the intravascular ultrasound image device of claim 1, wherein the further PROGRAMMED AUTO of processor monitors the variation of detected ultrasonic signal, purpose is to discern it for image intensification and body cavity at least

21. according to the intravascular ultrasound image device of claim 20, processing wherein comprises one of the transient change of difference tissue and instantaneous filtering at least.

22. according to the intravascular ultrasound image device of claim 1, further PROGRAMMED AUTO monitoring cardiovascular cycle of processor wherein.

23. according to the intravascular ultrasound image device of claim 1, wherein the further PROGRAMMED AUTO of processor monitors image quality.

24. an intravascular ultrasound image device comprises

Place endoceliac ultrasonic signal generator and detector; And

With the processor that ultrasonic signal generator links to each other with detector, this processor is program control to be finished:

A. by deriving first image in the first serial ultrasonic signal that detects,

B. by deriving second image in the second series ultrasonic signal that detects,

C. compare first image and second image,

D. monitor the variation of the ultrasonic signal that detects automatically,

E. the automatic monitoring cardiovascular cycle, and

F. relative second image stability, first image.

25. an intravascular ultrasound image device comprises

Ultrasonic signal generator and detector place in the body cavity and pass in the cross section of body cavity mobile;

With the processor that ultrasonic signal generator links to each other with detector, this processor is program control to be finished:

A. pass in the moving process first time of described cross section at ultrasonic signal generator and detector, by derivation first image in the ultrasonic signal that detects,

B. pass in the moving process second time of described cross section at ultrasonic signal generator and detector, by derivation second image in the ultrasonic signal that detects,

C. compare first image and second image, and

D. handle first image and second image; And

The display that links to each other with processor, processor are wherein adjusted the demonstration of second image according to result relatively.

26. an intravascular ultrasound image device comprises

Ultrasonic signal generator and detector place in the body cavity and move in the cross section of passing body cavity;

With the processor that ultrasonic signal generator links to each other with detector, this processor is program control to be finished:

A. by from the detected ultrasonic signal of the first in described cross section, deriving first image,

B. by from the detected ultrasonic signal of the second portion in described cross section, deriving second image,

C. compare first image and second image, and

D. handle first image and second image; And

The display that links to each other with processor, processor are wherein adjusted the demonstration of second image according to result relatively.

27. an intravascular ultrasound image device comprises

Place endoceliac ultrasonic signal generator and detector; And

With the processor that ultrasonic signal generator links to each other with detector, this processor is program control to be finished:

A. by deriving first image in the first serial ultrasonic signal that detects,

B. by deriving second image in the second series ultrasonic signal that detects,

C. monitor automatically, and

D. relative first image is judged second image.

28. according to the intravascular ultrasound image device of claim 27, processor wherein monitors that automatically first image and second image are so that observation vasodilation.

29. according to the intravascular ultrasound image device of claim 28, vasodilation wherein is one of local vascular dilation and overall vasodilation at least.

30. according to the intravascular ultrasound image device of claim 27, the further program control formation approximating function of processor wherein.

31. according to the intravascular ultrasound image device of claim 30, approximating function wherein utilizes one of cross-correlation, normalized crosscorrelation and SAD formation at least.

32. according to the intravascular ultrasound image device of claim 30, processor wherein monitors approximating function automatically so that the observation cardiovascular cycle.

33. according to the intravascular ultrasound image device of claim 32, processor wherein utilizes at least that threshold value is relevant, one of self correlation, Fourier transformation and spectrum analysis monitor approximating function automatically, so that the observation cardiovascular cycle.

34., wherein analyze approximating function so that determine the quality of image according to the intravascular ultrasound image device of claim 30.

35. according to the intravascular ultrasound image device of claim 30, evaluation wherein comprises the displacement judgement.

36. an intravascular ultrasound image device comprises

Place endoceliac ultrasonic signal generator and detector; And

With the processor that ultrasonic signal generator links to each other with detector, this processor is program control to be finished:

A. by deriving first image in the first serial ultrasonic signal that detects,

B. by deriving second image in the second series ultrasonic signal that detects,

C. relative first image is judged second image, and

D. relative first image stability, second image.

37. according to the intravascular ultrasound image device of claim 36, wherein further comprise the display that links to each other with described processor, second image after this display is used to show first image and stablize.

38. according to the intravascular ultrasound image device of claim 36, wherein relative first image stability, second image utilizes one of Cartesian coordinate and polar coordinate to finish at least.

39. according to the intravascular ultrasound image device of claim 36, wherein relative first image stability, second image carries out on a size at least.

40. according to the intravascular ultrasound image device of claim 36, wherein stable comprises at least to be stablized the displacement of flute card, rotatablely moves and one of vasodilation.

41. according to the intravascular ultrasound image device of claim 40, wherein stable comprises the stable overall situation, local and regular one of move at least.

42. according to the intravascular ultrasound image device of claim 36, stablize each in a plurality of positions in second image stable comprising wherein.

43. according to the intravascular ultrasound image device of claim 36, stable displacement second image that comprises wherein.

44. according to the intravascular ultrasound image device of claim 36, adjust second image stable comprising according to the result who judges wherein.

45. according to the intravascular ultrasound image device of claim 36, processor wherein is further program control for limiting drift.

46. according to the intravascular ultrasound image device of claim 43, processor wherein is further program control to be the restriction drift, described restriction drift realizes by utilizing the information adjustment that is obtained by the supervision of cardiovascular cycle to be shifted.

47. an intravascular ultrasound image formation method comprises the steps:

Place ultrasonic signal generator and detector in body cavity;

Detect ultrasonic signal;

By deriving first image in the ultrasonic signal that detects;

By deriving second image in the ultrasonic signal that detects;

Compare first image and second image; And

Handle first image and second image.

48., further comprise the step that shows first image and second image according to the intravascular ultrasound image formation method of claim 47.

49., judge second image comprising relative first image according to the intravascular ultrasound image formation method of claim 47.

50., wherein derive image and comprise one of processing and digitized at least according to the intravascular ultrasound image formation method of claim 47.

51., wherein comprise the formation two-dimensional array to going out image according to the intravascular ultrasound image formation method of claim 47.

52. according to the intravascular ultrasound image formation method of claim 51, wherein said two-dimensional array is to form in one of polar coordinate system and Cartesian coordinates at least.

53. according to the intravascular ultrasound image formation method of claim 51, wherein said two-dimensional array has a plurality of elements, each representative in a plurality of elements is from the predetermined detected ultrasonic signal in locus.

54. according to the intravascular ultrasound image formation method of claim 49, judgement wherein comprises the displacement judgement.

55. according to the intravascular ultrasound image formation method of claim 49, judgement wherein comprises and at least once approaches operation.

56. according to the intravascular ultrasound image formation method of claim 55, the operation that at least once approaches wherein comprises one of cross-correlation, normalized crosscorrelation and SAD at least.

57. according to the intravascular ultrasound image formation method of claim 56, cross-correlation wherein comprises one of direct cross-correlation and Fourier transformation at least.

58. according to the intravascular ultrasound image formation method of claim 49, judgement wherein utilizes one of polar coordinate system and Cartesian coordinates to finish at least.

59. according to the intravascular ultrasound image formation method of claim 49, judgement is wherein carried out on a size at least.

60., further comprise the step that detects flute card displacement at least, rotates one of mobile and vasodilation according to the intravascular ultrasound image formation method of claim 47.

61. according to the intravascular ultrasound image formation method of claim 60, wherein at least displacement of flute card and rotation one of to move be regular.

62. according to the intravascular ultrasound image formation method of claim 60, wherein at least displacement of flute card and rotation one of to move be partial.

63. according to the intravascular ultrasound image formation method of claim 60, wherein vasodilation is overall.

64. according to the intravascular ultrasound image formation method of claim 60, wherein vasodilation is partial.

65., further comprise the step that the detected ultrasonic signal of automatic supervision changes according to the intravascular ultrasound image formation method of claim 47.

66., further comprise the step of image intensification according to the intravascular ultrasound image formation method of claim 65.

67., further comprise the step of body cavity identification according to the intravascular ultrasound image formation method of claim 65.

68., further comprise the step in automatic monitoring cardiovascular cycle according to the intravascular ultrasound image formation method of claim 47.

69., further comprise the step of automatic supervision image quality according to the intravascular ultrasound image formation method of claim 47.

70. an intravascular ultrasound image formation method comprises the following steps:

Place ultrasonic signal generator and detector in body cavity;

Detect ultrasonic signal;

By deriving first image in the first serial ultrasonic signal that detects;

By deriving second image in the second series ultrasonic signal that detects;

Automatically monitor;

Relative first image is judged second image; And

Handle first image and second image.

71. the intravascular ultrasound image formation method according to claim 70 further comprises the step that forms approximating function.

72. according to the intravascular ultrasound image formation method of claim 71, approximating function wherein utilizes one of cross-correlation, normalized crosscorrelation and SAD formation at least.

73. according to the intravascular ultrasound image formation method of claim 70, automatic supervision wherein is to monitor first image and second image, so that observation vasodilation.

74. according to the intravascular ultrasound image formation method of claim 73, vasodilation wherein is one of local vascular dilation and overall vasodilation at least.

75. according to the intravascular ultrasound image formation method of claim 71, automatic supervision wherein is to monitor approximating function, so that the observation cardiovascular cycle.

76. according to the intravascular ultrasound image formation method of claim 75, automatic supervision wherein comprises at least that threshold value is relevant, one of self correlation, Fourier transformation and spectrum analysis.

77. according to the intravascular ultrasound image formation method of claim 71, comprising analyzing approximating function so that determine the quality of image.

78. according to the intravascular ultrasound image formation method of claim 71, judgement wherein comprises the displacement judgement.

79. an intravascular ultrasound image formation method comprises the following steps:

Place ultrasonic signal generator and detector in body cavity;

Detect ultrasonic signal;

By deriving first image in the first serial ultrasonic signal that detects;

By deriving second image in the second series ultrasonic signal that detects;

Automatically monitor;

Relative first image is judged second image; And

Relative first image stability, second image.

80., wherein further comprise the step of second image after showing first image and stablizing according to the intravascular ultrasound image formation method of claim 79.

81. according to the intravascular ultrasound image formation method of claim 79, stablizing wherein utilizes one of Cartesian coordinate and polar coordinate to finish at least.

82. according to the intravascular ultrasound image formation method of claim 79, stablizing wherein carried out on a size at least.

83. according to the intravascular ultrasound image formation method of claim 79, wherein stable comprises to be stablized the displacement of flute card at least, rotatablely moves and one of vasodilation.

84. 3 intravascular ultrasound image formation method according to Claim 8, wherein stable comprise the stable at least overall situation, local and regular move it.

85. 3 intravascular ultrasound image formation method according to Claim 8, stablize each in a plurality of positions in second image stable comprising wherein.

86. according to the intravascular ultrasound image formation method of claim 79, stable displacement second image that comprises wherein.

87. according to the intravascular ultrasound image formation method of claim 79, adjust second image stable comprising according to the result who judges wherein.

88., further comprise the step of restriction drift according to the intravascular ultrasound image formation method of claim 79.

89. 6 intravascular ultrasound image formation method according to Claim 8 further comprises the step of restriction drift, and restricted passage utilization wherein monitors by the cardiovascular cycle that the information adjustment that obtains is shifted and realizes.

90. an intravascular ultrasound image formation method comprises the following steps:

Place ultrasonic signal generator and detector in body cavity;

Detect ultrasonic signal;

By deriving first image in the first serial ultrasonic signal that detects;

By deriving second image in the second series ultrasonic signal that detects;

Compare first image and second image;

Automatically monitor the variation of the ultrasonic signal that detects;

The automatic monitoring cardiovascular cycle; And

Relative first image stability, second image.

91. an intravascular ultrasound image formation method comprises the following steps:

Place ultrasonic signal generator and detector in body cavity;

Ultrasonic signal generator and detector pass the cross section of body cavity and move;

Detect ultrasonic signal;

Derive first image in the detected ultrasonic signal in the moving process for the first time by pass described cross section at ultrasonic signal generator and detector;

Derive second image in the detected ultrasonic signal in the moving process for the second time by pass described cross section at ultrasonic signal generator and detector;

Compare first image and second image;

Adjust second image; And

Show second image of adjusting.

92. an intravascular ultrasound image formation method comprises the following steps:

Place ultrasonic signal generator and detector in body cavity;

Ultrasonic signal generator and detector pass the cross section of body cavity and move;

Detect ultrasonic signal;

By derivation first image the ultrasonic signal that detects from the first in described cross section,

By deriving second image the ultrasonic signal that detects from the second portion in described cross section,

Compare first image and second image;

Adjust second image; And

Show second image of adjusting.

93. an intravascular ultrasound image formation method comprises the following steps:

Place ultrasonic signal generator and detector in body cavity;

Detect ultrasonic signal;

By deriving first sequence image in the first serial ultrasonic signal that detects;

By deriving second sequence image in the second series ultrasonic signal that detects;

Compare first sequence image and second sequence image; And

Automatically mate first sequence image and second sequence image.

94. according to the intravascular ultrasound image formation method of claim 93, coupling wherein comprises the corresponding image of identification.

95. according to the intravascular ultrasound image formation method of claim 93, wherein the part of at least the first sequence image is the reference field part, and the part of at least the second sequence image is non-reference field part.

96. according to the intravascular ultrasound image formation method of claim 95, wherein coupling comprises non-reference field part relative reference section transposition of partial one images.

97. according to the intravascular ultrasound image formation method of claim 95, wherein coupling comprises the partially stabilized non-reference section of relative reference section.

98. according to the intravascular ultrasound image formation method of claim 97, stablizing respectively wherein carried out reference field part and non-reference field each corresponding image partly.

99. according to the intravascular ultrasound image formation method of claim 97, wherein stable carries out each corresponding image of first image sequence and second image sequence respectively.

100. according to the intravascular ultrasound image formation method of claim 93, wherein coupling comprises and approaches operation.

101. according to the intravascular ultrasound image formation method of claim 100, the operation that approaches wherein comprises one of cross-correlation and normalized crosscorrelation.

102. intravascular ultrasound image formation method according to claim 93, the derivation from ultrasonic signal generator and detector moved along the first time in first cross section of body cavity of the first series image wherein, the derivation from ultrasonic signal generator and detector moved along the second time in second cross section of body cavity of second series image wherein.

103. according to the intravascular ultrasound image formation method of claim 102, wherein body cavity first cross section and second cross section are roughly expanded simultaneously.

104., wherein relatively comprise the relative first image judgement, second image according to the intravascular ultrasound image formation method of claim 93.

105. an intravascular ultrasound image device comprises:

At endoceliac ultrasonic signal generator and detector; And

With the processor that ultrasonic signal generator links to each other with detector, this processor is program control to be finished:

A. by deriving first sequence image in the first serial ultrasonic signal that detects;

B. by deriving second sequence image in the second series ultrasonic signal that detects;

C. compare first sequence image and second sequence image; And

D. mate first sequence image and second sequence image automatically.

106. according to the intravascular ultrasound image device of claim 105, wherein relatively first image comprises the relative first image judgement, second image with second image.

107. according to the intravascular ultrasound image device of claim 105, wherein coupling comprises the corresponding image of identification.

108. according to the intravascular ultrasound image device of claim 105, wherein the part of at least the first sequence image is the reference field part, and the part of at least the second sequence image is non-reference field part.

109. according to the intravascular ultrasound image device of claim 108, wherein coupling comprises non-reference field part relative reference section transposition of partial one images.

110. according to the intravascular ultrasound image device of claim 108, wherein coupling comprises the partially stabilized non-reference field part of relative reference section.

111. according to the intravascular ultrasound image device of claim 110, stablizing respectively wherein carried out reference field part and non-reference field each corresponding image partly.

112. according to the intravascular ultrasound image device of claim 110, wherein stable carries out each corresponding image of first image sequence and second image sequence respectively.

113. according to the intravascular ultrasound image device of claim 105, wherein coupling comprises and approaches operation.

114. according to the intravascular ultrasound image device of claim 113, the operation that approaches wherein comprises one of cross-correlation and normalized crosscorrelation.

115. intravascular ultrasound image device according to claim 105, the derivation from ultrasonic signal generator and detector moved along the first time in first cross section of body cavity of the first series image wherein, the derivation from ultrasonic signal generator and detector moved along the second time in second cross section of body cavity of second series image wherein.

116. according to the intravascular ultrasound image device of claim 115, body cavity first cross section wherein and second cross section roughly are diastoles simultaneously.

117. the intravascular ultrasound image device according to claim 1 further comprises the probe that links to each other with detector with ultrasonic signal generator.

118. according to the intravascular ultrasound image device of claim 117, probe wherein is one of conduit and lead at least.

119. according to the intravascular ultrasound image device of claim 1, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

120. the intravascular ultrasound image device according to claim 25 further comprises the probe that links to each other with detector with ultrasonic signal generator, and passes the cross section and move ultrasonic signal generator and detector.

121. according to the intravascular ultrasound image device of claim 120, probe wherein is one of conduit and lead at least.

122. according to the intravascular ultrasound image device of claim 25, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

123. the intravascular ultrasound image device according to claim 26 further comprises the probe that links to each other with detector with ultrasonic signal generator, and passes the cross section and move ultrasonic signal generator and detector.

124. according to the intravascular ultrasound image device of claim 123, probe wherein is one of conduit and lead at least.

125. according to the intravascular ultrasound image device of claim 26, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

126. according to the intravascular ultrasound image device of claim 27, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

127. according to the intravascular ultrasound image device of claim 36, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

128. according to the intravascular ultrasound image device of claim 47, ultrasonic signal generator wherein links to each other with probe with detector.

129. according to the intravascular ultrasound image device of claim 128, probe wherein is one of conduit and lead at least.

130. according to the intravascular ultrasound image device of claim 47, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

131. according to the intravascular ultrasound image device of claim 70, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

132. according to the intravascular ultrasound image device of claim 79, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

133. according to the intravascular ultrasound image device of claim 91, wherein ultrasonic signal generator links to each other with probe with detector, this probe moves ultrasonic signal generator and detector.

134. according to the intravascular ultrasound image device of claim 133, probe wherein is one of conduit and lead at least.

135. according to the intravascular ultrasound image device of claim 91, wherein ultrasonic signal generator links to each other with probe with detector, this probe moves ultrasonic signal generator and detector.

136. according to the intravascular ultrasound image device of claim 92, wherein ultrasonic signal generator links to each other with probe with detector, this probe moves ultrasonic signal generator and detector.

137. according to the intravascular ultrasound image device of claim 136, probe wherein is one of conduit and lead at least.

138. according to the intravascular ultrasound image device of claim 92, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

139. according to the intravascular ultrasound image device of claim 102, wherein ultrasonic signal generator links to each other with probe with detector, this probe moves ultrasonic signal generator and detector.

140. according to the intravascular ultrasound image device of claim 139, probe wherein is one of conduit and lead at least.

141. according to the intravascular ultrasound image device of claim 93, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

142. according to the intravascular ultrasound image device of claim 115, comprise that further probe links to each other with detector with ultrasonic signal generator, and in the cross section mobile ultrasonic signal generator and detector.

143. according to the intravascular ultrasound image device of claim 142, probe wherein is one of conduit and lead at least.

144. according to the intravascular ultrasound image device of claim 105, ultrasonic signal generator wherein and detector comprise independently ultrasonic signal generator and one detector independently.

Images