US20020012417A1

US20020012417A1 - Method for reducing line artifacts in a CT image and device for implementing the method

Info

Publication number: US20020012417A1
Application number: US09/797,330
Authority: US
Inventors: Herbert Bruder; Thomas Flohr; Karl Stierstorfer; Thomas Haar
Original assignee: Individual
Current assignee: Siemens AG
Priority date: 2000-03-01
Filing date: 2001-03-01
Publication date: 2002-01-31
Anticipated expiration: 2021-03-01
Also published as: US6411671B2; DE10009746B4; DE10009746A1; JP2001276055A

Abstract

In a method and apparatus for reducing line artifacts in a CT image D₁, which is acquired during a scan by a CT device with a detector system having a number of proper detector channels and at least one defective detector channel, the image M₁is obtained first by a median filtering. the image D₁, which contains a circle artifact on a circle K₁. A difference value image F₁=D₁−M₁is then produced. Filtering then is carried out in each picture element of the image F₁in the direction of the tangents t¹ ₁and t² ₁to the circle K₁extending through the respective picture element in order to produce two resulting images G¹ ₁and G² ₁. A correction image D₂is subsequently produced by subtracting the resulting images G¹ ₁and G² ₁from D₁.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention is directed to a method and apparatus for reducing line artifacts in a CT image, wherein the CT image is produced by scanning with an X-ray source which is rotatable around an examination subject, with X-rays from the X-ray source, after being attenuated by an examination subject, being incident on a detector system.

2. Description of the Prior Art

CT devices are known which have an X-ray source, e.g. an X-ray tube, which direct a collimated, pyramid-shaped X-ray bundle through the examination subject, e.g. a patient, onto a detector system that is composed of a number of detector channels. Each detector channel has at least one detector element and one associated electronic element for reading out and amplifying the signal that is generated in the detector element as a result of the incident radiation. A number of detector elements can be allocated to one electronic element. The X-ray source and, depending on the construction of the CT device, the detector system as well are attached to a gantry that rotates around the examination subject. A support device for the examination subject can be displaced along the system axis relative to the gantry. The position along the system axis at which the X-ray bundle penetrates the examination subject, and the angle, under which the X-ray bundle penetrates the examination subject, are continuously modified as a result of the displacement and the rotation of the gantry. Each detector element of the detector system struck by the radiation produces a signal representing a measure of the overall transparency of the examination subject for the radiation proceeding from the X-ray source to the detector system. The set of output signals of the detector elements of the detector system, which is acquired for a specific position of the X-ray source, is referred to as a projection. A scan is composed of a set of projections, which are acquired at different positions of the gantry and/or at different positions of the support device. The CT device picks up a number of projections during a scan in order to be able to construct a two-dimensional tomogram of a slice of the examination subject. A number of slices can be picked up at the same time by a detector system that is composed of an array having a number of rows and columns of detector elements. Such planar-like detector systems, however, frequently contain detector channels which do not supply proper data. It may be that detector system contains faulty detector channels already after the production process, for example as a result of defects in fabrication caused by the high integration density of the electronic elements. Defects of individual detector channels also may arise during the operation of the CT device. Such defects cause circular structures in the acquired CT images, these circular structures being referred to as circle artifacts. Techniques referred to as “ring-balancing” methods are known from the literature for the purpose of attempting to correct or prevent such artifacts in CT images. Such methods are disclosed in U.S. Pat. No. 4,670,840 and in German OS 198 35 451 (corresponding to U.S. Pat. No. 6,047,039), for example.

A disadvantage of such known methods is that they insufficiently eliminate artifacts which arise in a CT device having a defective detector channel.

SUMMARY OF THE INVENTION

An object of the present invention is to provide a method for reducing artifacts in a CT image, so that the obtainable image quality is improved in a CT device having at least one defective detector channel. It is also an object of the invention to provide a CT device for implementing the method.

The above object is achieved in accordance with the principles of the present invention in a method for reducing line artifacts in a CT image D ₁, as well as in an apparatus for implementing the method, wherein the image has been subjected to interpolated filtering for preventing or correcting faulty values of picture elements represented in a circle K₁, and wherein the following steps are implemented. The picture elements of the CT image D₁are subjected to a median filtering, orthogonal to the straight line extending through the respective picture element and the center of the circle K₁, for producing an image M₁. A difference value image F₁is generated by subtracting the image M₁from the CT image D₁. Two resulting images G¹ ₁and G² ₁are produced by filtering the picture elements of the difference value image F₁in the respective directions of tangents t¹ ₁and t² ₁to the circle K₁extending through the respective picture element. Filtering is conducted along t¹ ₁to produce the resulting image G¹ ₁and is conducted along t² ₁for producing the resulting image G² ₁. A correction image D₂is then obtained by subtracting both of the resulting images G¹ ₁and G² ₁from the image D₁.

Defective detector channels of a detector system lead to faulty values for picture elements, which appear as a circle in a CT image acquired by the detector system. Such image errors therefore are referred to as circle artifacts. Defective detector channels are not only ones that fail to supply an output signal as a result of the defect, but also are channels with a measuring accuracy that exceeds a specific tolerance value. Various methods are known for correcting or preventing circle artifacts, as noted above. These methods are carried out on the measurement data or on the image data and are primarily based on interpolated filtering, and achieve a noticeable weakening of the circle artifacts in the acquired CT images. Such known methods have the disadvantage, however, that line-like image errors frequently arise in the resulting CT images after such a method has been implemented. These image errors are referred to as line artifacts. They increasingly occur in association with large signal unsteadiness caused by high-contrast areas of an examination subject. The inventive method is particularly advantageous for eliminating such line artifacts, which occur after circle artifacts have been eliminated and which appear as tangents to the circles in the CT image. The size and position data of the circles of the circle artifacts caused by the defective detector channels are assumed to be known. It is sufficient to know the position of the circle center, which is the same for all circles, and the radii of the circles.

The elimination of the line artifacts is initially described for the case of a single defective detector channel. The faulty values of picture elements caused by the defective detector channel are situated on a circle K ₁. These faulty values are corrected by a known ring-balancing method on the measuring data or on the image data. Line artifacts, which appear as tangents to the circle K₁, arise in the resulting CT image D₁. For producing an image M₁, a median filtering is carried out for each picture element of the CT image D₁situated outside of the circle K₁, orthogonally to the straight line extending through the respective picture element and the center of the circle K₁. The width of the median filtering can be modifiable. This width should be selected wider than the expected line width of the line artifacts. A median filter of the width 5 has proven to be beneficial. The sampling distance A₁must be selected dependent on the convolution kernel used for the reconstruction. Ideally, the image M₁no longer contains line artifacts.

A difference value image F ₁=D₁−M₁is generated by subtracting the image M₁from the CT image D₁. For producing two resulting images G¹ ₁and G² ₁, a filtering is carried out in each picture element of the difference value image F₁in the respective directions of the tangents t¹ ₁and t² ₁to the circle K₁extending through the picture element in question, i.e., the filtering is carried out along t¹ ₁for the resulting image G¹ ₁and the filtering is carried out along t² ₁for the resulting image G² ₁. This step is to eliminate, for the most part, the image pixel noise difference value in the image F₁for the most part. Furthermore, the line artifacts that are present in the error image F₁are emphasized in the result images. It is necessary to calculate two result images G¹ ₁and G² ₁, since there are two possible tangent directions to the circle K₁per picture element. The thus-determined resulting images G¹ ₁and G² ₁are subtracted from D₁and the correction image D₂is obtained, which ideally no longer contains line artifacts generated by the defective channel.

The correction image D ₂serves as basis image for correcting the line artifacts caused by other detector channels if the detector system contains further defective detector channels. In general, the correction image D₁serves as a basis image for correcting line artifacts caused by the i-th defective detector channel. Since the image M₁has been calculated in relation to the center of the circle that is valid for all circles k_i, the difference value image F₁can still be used in the following. For producing the result images G¹ _iand G² _i, a filtering is carried out in each picture element of the difference value image F₁situated outside of the circle K_iin the direction of the tangents D¹ _iand D² _ito the circle K_i, i.e., the filtering is carried out along D¹ _ifor the resulting image G¹ _iand the filtering is carried out along D² _ifor the resulting image G² _i. The correction image D_i−1=D_i−G¹ _i−G² _iis determined by subtracting the resulting images G¹ _iand G² _ifrom D_i. This procedure is repeated until the line artifacts of all defective detector channels are eliminated.

The above-described method is relatively time-consuming as a result of the serial processing of the line artifacts caused by the defective detector channels. Parallel processing of the image errors caused by the different defective detector channels is preferable for processing the data faster. Accordingly, a further embodiment of the inventive method proceeds as described above until the calculation of the difference value image F ₁. The resulting images G¹ _iand G² _iare calculated in this manner for all defective detector channels. In contrast to the initially described embodiment, different correction images are not consecutively determined, which respectively serve as an image for calculating the next correction images, but instead all resulting images G¹ _iand G² ₁are subtracted from D₁. The correction image D₂results. The defective detector channels should have a minimum distance (spacing) of ten detector channels as a condition for this parallel processing.

The aforementioned median filtering for producing an image M ₁, in each picture element of the initial CT image, takes place orthogonally to the straight line extending through the picture element in question and the center of the circle K_i. This calculation represents an approximation, and it is an advantage of the inventive method that it must be carried out only once for each picture element even given a number of defective detector channels. Better results are obtained, however, by the median filtering orthogonal to the tangents to the circles K_i. This makes it necessary to carry out the pixel-oriented median filtering anew for each defective detector channel, with an associated high computing outlay. Given a single defective detector channel and the faulty values of picture elements on a circle K₁caused as a result thereof, the correction includes the following steps:

carrying out an interpolated filtering of the artifact-containing CT image, thereby leading to the CT image D ₁, in order to avoid or to correct faulty values of picture elements on the circle K₁,

carrying out a median filtering in each picture element of the CT image D ₁situated outside of the circle K₁, orthogonal to the tangents to the circle K₁extending through the respective picture element, for producing an image M₁,

generating a difference value image F ₁D₁−M₁by subtracting the image M₁from the CT image D₁,

carrying out a filtering in each picture element of the difference value image F ₁situated outside of the circle K₁in the direction of the tangents t¹ ₁and t² ₁to the circle K₁extending through the respective picture element in order to produce two resulting images G¹ ₁and G² ₁, with the filtering being carried out along t¹ ₁for the resulting image G¹ ₁and the filtering is carried out along t² ₁for the resulting image G² ₁.

determining the correction image D ₂=D₁−G¹ ₁−G² ₁by subtracting the resulting images G¹ ₁and G² ₁are from D₁.

If a number of defective detector channels are present, the following applies for processing the i-th defective detector channel:

a median filtering is carried out in each picture element of the CT image D _i, orthogonal to the tangents to the circle K_iextending through the respective picture element, for producing an image M_i,

a difference value image F _i=D_i−M_iis produced by subtracting the image M_ifrom the correction image D_i,

a filtering is carried out in each picture element of the difference value image F ₁situated outside of the circle K₁in the direction of the tangents t¹ _iand t² _ito the circle K_iextending through the respective picture element in order to produce two resulting images G¹ _iand G² _i, with the filtering being carried out along t¹ _ifor the resulting image G¹ _iand the filtering being carried out along t² _ifor the resulting image G² _i.

the correction image D _i−1=D_i−D_i−G¹ _i−G² _iis determined by subtracting the resulting images G¹ _iand G² _ifrom D_i.

These steps are repeated until all image errors caused by the different defective detector channels are processed.

In a further version of the invention, a running averaging serves as the filtering for producing the resulting images G ¹ _iand G² _i. In another version of the invention, a sum operator is implemented as the filtering for producing the resulting images G¹ _iand G² _i. A combination of these two versions also can be employed, wherein a running averaging and a sum operator are simultaneously employed for filtering.

At the beginning of each of the aforementioned embodiments, the values of the CT image D ₁are preferably limited to a range 1000 HU (H₂O)±Δ. This limitation is expedient for examining soft-tissue parts, in which disturbances caused by defective detector channels are particularly apparent in the image.

The intensity of the noise signal in the resulting images G ¹ _iand G² ⁱis dependent on the filter width of the running averaging or of the sum operator. This noise signal influences the correction images D₁₊₁and therefore can lead to undesired noise structures in the resulting image in the case of a number of defective detector channels. A high filter width is required as a result. The intensity of a line artifact varies, however, so that a limitation of the filter width of the running averaging or of the sum operator is required.

The summation of the HU values in G ¹ _iin the tangent direction D¹ _ior in G² _iin the tangent direction D² ⁱrepresents an additional measure for suppressing noise. This corresponds to the calculation of the radon values in these tangent directions. This operation of the image processing is referred to as HUG transformation. On the basis of a threshold criterion, a noise signal can be principally differentiated from artifact structures and can be eliminated. This threshold must be suitably selected, however, so that low-contrast lines in G¹ _iand G² _i, that are actually a part of the diagnostically relevant image context, are not erroneously eliminated.

DESCRIPTION OF THE DRAWINGS

FIG. 1 schematically illustrates the basic components of an X-ray computed tomography apparatus, constructed and operating in accordance with the invention. [0029]
FIG. 2 is a flow chart for describing an exemplary embodiment of the inventive method for reducing line artifacts. [0030]
FIG. 3 is a schematic representation of a median filtering in the inventive method. [0031]
FIG. 4 is a schematic representation of a running averaging in the inventive method. [0032]
FIG. 5 illustrates the extension of the method according to FIG. 2 given more than one defective detector channel. [0033]
FIG. 6 is a flow chart for the inventive method for reducing line artifacts with an exact calculation of error images. [0034]
FIG. 7 is a flow chart for the inventive method for reducing line artifacts with some of the data processing taking place in parallel. [0035]
FIG. 8 shows a reference CT image of the skull base. [0036]
FIG. 9 shows a CT image having high-contrast line artifacts. [0037]
FIG. 10 shows the CT image of FIG. 9 after median filtering in accordance with the inventive method. [0038]
FIG. 11 shows the difference value image formed from the respective images of FIG. 10 and FIG. 9 in accordance with the inventive method. [0039]
FIG. 12 shows the resulting image G[0040] ¹ ₁formed by a running averaging of the difference value image of FIG. 11 in accordance with the inventive method.
FIG. 13 shows the resulting image G[0041] ² ₁formed by a running averaging of the difference value image of FIG. 11 in accordance with the inventive method.
FIG. 14 shows the correction image D[0042] 2 obtained in accordance with the inventive method.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

FIG. 1 shows a CT device, which is provided for scanning an [0043] examination subject 1, having an X-ray source 2, such as an X-ray tube, with a focus 3 from which a pyramid-shaped X-ray bundle 4 is emitted, which is gated by a radiation diaphragm (not shown) and which penetrates the examination subject 1, for example a patient, and strikes a detector system 5. The detector system 5 is an array of parallel rows 6 and parallel columns 7 of detector elements 8. The X-ray source 2 and the detector system 5 form a measuring system, which can be rotated in a direction 6 around a system axis 9 and which can be displaced along the system axis 9 relative to the examination subject 1, so that the examination subject 1 is irradiated from different projection angles and in different z-positions along the system axis 9. A signal processing unit 10 forms measurement values from the output signals of the detector elements 8 of the detector system 5. These measurement values are supplied to a computer 11, which calculates an image of the examination subject 1, which is reproduced at the monitor 12.
The flow chart as shown in FIG. 2 illustrates the inventive method for reducing line artifacts given a defective detector channel. The artifacts caused by the defective detector channel are situated on a circle K[0044] ₁. They are already corrected in a known manner in the CT image D₁, so that the circle K₁is no longer visible therein. As an undesired side effect of this correction, line artifacts, as tangents to the circle K₁, appear in the CT image D₁. A median filtering orthogonal to the straight line, which extends through the respective picture element of the CT image D₁and the center of the circle K₁, produces the image M₁from the CT image D₁. A difference value image F₁is produced by subtracting the image M₁from the CT image D₁. Two resulting images G¹ ₁and G² ₁are derived from the difference value image F₁as a result of a further filter operation. The filtering ensues on the basis of each picture element of F₁situated outside of the circle K₁in the direction of the tangents to K₁. The filtering is carried out along the tangent t¹ ₁for the resulting image G¹ ₁and is carried out along the tangent t² ₁for the resulting image G² ₁. The line artifacts that are present in the error image F₁thus are emphasized in the resulting images G¹ ₁, and G² ₁. The two resulting images G¹ ₁and G² ₁are subsequently subtracted from the original CT image D₁, and the CT image D₂is obtained, which no longer contains any visible line artifacts.
FIG. 3 schematically shows an exemplary median filtering. Shown are the circle K[0045] ₁with the center m and radius r, the two picture elements P₁and P₂, as well as the straight lines g₁and g₂, respectively through P₁or P₂and the center m. Furthermore, the positions of adjacent picture elements for P₁and P₂is indicated, which are situated on an orthogonal straight line with regard to g₁and g₂, respectively. Given the median filtering, a mean value is formed in the example from P₁and the four illustrated adjacent picture elements, and from P₂and the four corresponding adjacent picture elements, and is utilized as a new image value of P₁and P₂, respectively.
FIG. 4 schematically shows pixel-dependent averages. The directions of the averages in the picture element P[0046] ₁are fixed by the tangents t¹ ₁and t² ₁to the circle K₁having the radius r and the center m, this circle K₁being defined by the defective detector channel.
FIG. 5 shows the expansion of the method of FIG. 2 to more than one defective detector channel. On the basis of the difference value image F[0047] ₁, the two result images G¹ _iand G² _iare produced by filtering the picture elements of F₁in the direction of the tangents t¹ _i, and t² ₁to the circle K_iextending through the respective picture element. The filtering along the tangent t¹ _ileads to the resulting image G¹ _iand the filtering along the tangent t² _ileads to the resulting image G² _i. The resulting images G¹ ₁and G² ₁are subtracted from the image D_i. The image D_i+1arises. The index i is increased by 1 and the method is repeated until i>M. The last-determined image D_iwith i=M+1 is displayed as the CT image at the monitor.
In contrast to FIG. 5, the error image F[0048] _iis exactly calculated for each defective detector channel in the method shown in FIG. 6. For this purpose, the median filtering is carried out for each picture element of D_isituated outside of the circle K_iin the direction of the tangents to the circle K_iextending through the respective picture element. The number of error images to be calculated in this method therefore corresponds to the number of defective detector channels. In addition to the error image, the two resulting images G¹, and G² ₁therefore are produced anew for each detector channel. The remaining method steps corresponds to the method steps described in FIGS. 2 and 5. This method, which is more complicated compared to that of FIG. 2 or FIG. 5, also is repeated until all defective detector channels are processed.
FIG. 7 shows an exemplary method embodiment, wherein the resulting images G[0049] ¹ _iand G² _iare not sequentially calculated as in the previously described embodiments for a number of defective detector channels, but instead the result images G¹ _iand G² _i, 1≦i≦M, are calculated in parallel for the multiple defective detector channels. On the basis of the CT image D₁, the image M₁also initially arises as a result of a median filtering, and this image M₁is subtracted from D₁for calculating the image F₁. The resulting images G¹ _iand G² _i, 1≦i≦M, which are determined in parallel on the basis of F₁, also are subtracted from the CT image D₁for producing the CT image D₂. Assuming the availability of a very powerful computer, this method has a time advantage for the correction of the line artifacts in the CT images. Since the use of filter for calculating a resulting image has an effect on the adjacent resulting images, the defective detector elements should have a minimum spacing of ten channels from each other in order to be able to reconstruct a CT image that is free of line artifacts.
The efficacy of the method is demonstrated below on the basis of actual image data, acquired with a [0050] SIEMENS SOMATOM PLUS 4 VOLUME ZOOM CT device.
As a reference CT image, FIG. 8 shows image data of the skull basis given a scan carried out in a 4×1 slice modus, axial scan operation. The measuring data of a detector row are evaluated, whereby a predefined individual channel does not supply measurement values. A linear interpolation eliminates the circle artifacts arising as a result of the “defective” channel. As shown in FIG. 9, the result is the CT image, which serves as an initial CT image D[0051] ₁for the inventive method and which has a high-contrast line artifact at the location marked by the arrow, for example. The method as shown in FIG. 2 is used for eliminating the line artifacts. FIG. 10 shows the median-filtered input image M₁. Given the subtraction of M₁from the original CT image D₁, the difference value image F₁essentially showing the line artifacts and image noise results as shown in FIG. 11. A running averaging in the direction of the tangents of the circle caused by the defective detector channel subsequently results for each pixel of the difference value image F₁. Since—proceeding from each pixel of the difference value image F₁—there are two possible tangents to the circle, two resulting images G¹ ₁and G² ₁are produced. These are shown in the FIGS. 12 and 13. Ideally, the image noise is mainly suppressed in these images. When the two resulting images G¹ ₁, and G² ₁are subtracted from the original CT image D₁as shown in FIG. 9, the correction image D₂as shown in FIG. 14 results. As can be seen from D₂, the line artifacts have been effectively suppressed without increasing the noise contribution.
For the practical realization of a method according to the invention, only the corrected images are displayed at the monitor. The method process therefore proceeds essentially on the basis of computer-internal “image data”, which are not reconstructed as “images” at the monitor. [0052]
In the described exemplary embodiments, CT devices of the third generation are shown, i.e., the X-ray source and the detector rotate together around the system axis during the image generation. The invention also can be used with CT devices of the fourth generation, wherein only the X-ray source rotates and cooperates with a stationary detector ring. [0053]
The described exemplary embodiments relate to the medical application of inventive CT devices. The invention can also be applied outside of the medical field baggage inspection or material testing, for example. [0054]
Although modifications and changes may be suggested by those skilled in the art, it is the intention of the inventors to embody within the patent warranted hereon all changes and modifications as reasonably and properly come within the scope of their contribution to the art. [0055]

Claims

We claim as our invention:

1. A method for reducing line artifacts in a CT image D₁, acquired by scanning an examination subject by rotating at least a focus of an X-ray source around the examination subject and detecting radiation, attenuated by said examination subject, with a detector system which has a plurality of proper detector channels and at least one defective detector channel, and wherein said CT image D₁is subjected to an interpolated filtering for preventing or correcting faulty values of picture elements represented on a circle K₁, said faulty values being caused by the defective detector channel, comprising the steps of:

(a) subjecting picture elements of the CT image D₁to a median filtering, orthogonal to a straight line extending through the respective picture element and a center of the circle K₁, for producing an image M₁;

(b) forming a difference value image F=D₁−M₁by subtracting the image M₁from the CT image D₁;

(c) producing two resulting images G¹ ₁, and G² ₁by filtering picture elements of the difference value image F₁in respective directions of tangents t¹ ₁and t² ₁to the circle K₁extending through the respective picture element, whereby the filtering is conducted along t¹ ₁for the resulting image G¹ ₁and is conducted along t² ₁for the resulting image G² ₁; and

(d) forming a correction image D₂=D₁−G¹ ₁−G² ₁by subtracting the resulting images G¹ ₁and G² ₁from D₁.

2. A method for reducing line artifacts in a correction image D₂as claimed in claim 1, wherein the number of the defective detector channels equals M (M>1) and wherein i is a running index with i=2 . . . M, further comprising:

(e) producing said two resulting images G¹ _iand G² _iby filtering picture elements of the difference value image F₁in respective directions of tangents t¹ _iand t² _ito the circle K₁extending through the respective picture element, whereby the filtering is conducted along t¹ _ifor the resulting image G¹ _iand is conducted along t² _ifor the resulting image G² _i;

(f) forming a correction image D_i+1=D_i−G¹ _i−G² _iby subtracting the resulting images G¹ _iand G² _ifrom D_i.

(g) increasing the index i by one, and repeating steps (f) to (h) until i>M.

3. A method as claimed in claim 1 comprising adjusting a width of said median filtering.

4. A method as claimed in claim 1 wherein the step of filtering said picture elements to produce said resulting images G¹ ₁and G² ₁comprises conducting a running averaging along the respective tangents t¹ ₁and t² ₂.

5. A method as claimed in claim 1 wherein the step of filtering said picture elements to produce the resulting images G¹ ₁and G² ₁comprises employing a sum operator along said tangents t¹ ₁and t² ₁.

6. A method as claimed in claim 1 wherein the step of filtering said picture elements to produce the resulting images G¹ ₁and G² ₁comprises conducting a running averaging and employing a sum operator along said tangents t¹ ₁and t² ₁.

7. A method as claimed in claim 1 comprising limiting values of the picture elements of the CT image D₁to a range 1,000 HU(H₂O)±Δ.

8. A method for reducing line artifacts in a CT image D₁, acquired by scanning an examination subject by rotating at least a focus of an X-ray source around the examination subject and detecting radiation, attenuated by said examination subject, with a detector system which has a plurality of proper detector channels and at least one defective detector channel, and wherein said CT image D₁is subjected to an interpolated filtering for preventing or correcting faulty values of picture elements represented on a circle K₁, said faulty values being caused by an i^thdefective detector channel, said method comprising the steps of:

(a) subjecting picture elements of the CT image D₁to a median filtering, orthogonal to a straight line extending through the respective picture element and a center of the circle K_i, for producing an image M₁;

(b) forming a difference value image F₁=D₁−M₁by subtracting the image M₁from the CT image D₁;

(c) producing two resulting images G¹ _iand G² _iby filtering picture elements of the difference value image F₁in respective directions of tangents t¹ _iand t² _ito the circle K_iextending through the respective picture element, whereby the filtering is conducted along t¹ _ifor the resulting image G¹ _iand is conducted along t² _ifor the resulting image G² _i; and

(d) forming a correction image D₂=D₁−G¹ _i−G² _i, 1≦i≦M, by subtracting the resulting images G¹ _iand G² _ifrom D₁.

9. A method as claimed in claim 8 comprising adjusting a width of said median filtering.

10. A method as claimed in claim 8 wherein the step of filtering said picture elements to produce said resulting images G¹ _iand G² _icomprises conducting a running averaging along the respective tangents t¹ _iand t² _i.

11. A method as claimed in claim 8, wherein the step of filtering said picture elements to produce the resulting images G¹ ₁and G², comprises employing a sum operator along said tangents t¹ ₁and t² _i.

12. A method as claimed in claim 8 wherein the step of filtering said picture elements to produce the resulting images G¹ _iand G² _icomprises conducting a running averaging and employing a sum operator along said tangents t¹ _iand t² _i.

13. A method as claimed in claim 8 comprising limiting values of the picture elements of the CT image D₁to a range 1,000 HU(H₂O)±Δ.

14. A computed tomography apparatus comprising:

an X-ray source for scanning an examination subject by rotating at least a focus of said X-ray source around the examination subject;

a detector system for detecting radiation, attenuated by said examination subject, which has a plurality of proper detector channels and at least one defective detector channel said detector system emitting electrical signals representing data comprising a CT image D₁; and

a computer connected to said detector system wherein said CT image D₁is subjected to an interpolated filtering for preventing or correcting faulty values of picture elements represented on a circle K₁, said faulty values being caused by the defective detector channel, for reducing line artifacts in said CT image D₁, said computer subjecting picture elements of the CT image D₁to a median filtering, orthogonal to a straight line extending through the respective picture element and a center of the circle K₁, for producing an image M₁, and forming a difference value image F₁=D₁−M₁by subtracting the image M₁from the CT image D₁, and producing two resulting images G¹ ₁, and G² ₁, by filtering picture elements of the difference value image F₁in respective directions of tangents t¹ ₁and t² ₁to the circle K₁extending through the respective picture element, whereby the filtering is conducted along t¹ ₁for the resulting image G¹ ₁and is conducted along t² ₁for the resulting image G² ₁, and forming a correction image D₂=D₁−G¹ ₁−G² ₁by subtracting the resulting images G¹ ₁and G² ₁from D₁.

15. A computed tomography apparatus as claimed in claim 14, wherein the number of the defective detector channels equals M (M>1) and wherein i is a running index with i=2 . . . M, wherein said computer produces said two result images G¹ _iand G² _iby filtering picture elements of the difference value image F₁in respective directions of tangents t¹ _iand t² _ito the circle K₁extending through the respective picture element, said computer conducting the filtering along t¹ ₁for the resulting image G¹ _iand along t² _ifor the resulting image G² ₁.

16. A computed tomography apparatus as claimed in claim 14 wherein said computer adjusts a width of said median filtering.

17. A computed tomography apparatus as claimed in claim 14 wherein said computer filters said picture elements to produce said resulting images G¹ ₁, and G² ₁by conducting a running averaging along the respective tangents t¹ ₁and t² ₂.

18. A computed tomography apparatus as claimed in claim 14 wherein said computer filters said picture elements to produce the resulting images G¹ ₁and G² ₁by employing a sum operator along said tangents t¹ ₁and t² ₁.

19. A computed tomography apparatus as claimed in claim 14 wherein said computer filters said picture elements to produce the resulting images G¹ ₁and G² ₁by conducting a running averaging and employing a sum operator along said tangents t¹ ₁and t² ₁.

20. A computed tomography apparatus as claimed in claim 14 wherein said computer limits values of the picture elements of the CT image D₁to a range 1,000 HU(H₂O)±Δ.

21. A computed tomography apparatus comprising:

a computer connected to said detector system wherein said CT image D₁is subjected to an interpolated filtering for preventing or correcting faulty values of picture elements represented on a circle K₁, said faulty values being caused by an i^thdefective detector channel, for reducing line artifacts in said CT image, said computer subjecting picture elements of the CT image D₁to a median filtering, orthogonal to a straight line extending through the respective picture element and a center of the circle K_i, for producing an image M₁, and forming a difference value image F₁=D₁−M₁by subtracting the image M₁from the CT image D₁, and producing two resulting images G¹ ₁and G² _iby filtering picture elements of the difference value image F₁in respective directions of tangents t¹ _iand t² _ito the circle K_iextending through the respective picture element, whereby the filtering is conducted along t¹ ₁for the resulting image G¹ _iand is conducted along t² _ifor the resulting image G² _i, and forming a correction image D₂=D₁−G¹ ₁−G² _i, 1≦i≦M, by subtracting the resulting images G¹ _iand G² _ifrom D₁.

22. A computed tomography apparatus as claimed in claim 21 wherein said computer adjusts a width of said median filtering.

23. A computed tomography apparatus as claimed in claim 21 said computer filters said picture elements to produce said resulting images G¹ _iand G² _icomprises conducting a running averaging along the respective tangents t¹ _iand t².

24. A computed tomography apparatus as claimed in claim 21 wherein said computer filters said picture elements to produce the resulting images G¹ _iand G² _icomprises employing a sum operator along said tangents t¹ _iand t² _i.

25. A computed tomography apparatus as claimed in claim 21 wherein said computer filters said picture elements to produce the resulting images G¹ _iand G² _iby conducting a running averaging and employing a sum operator along said tangents t¹ _iand t² _i.

26. A computed tomography apparatus as claimed in claim 21 wherein said computer limits values of the picture elements of the CT image D₁to a range 1,000 HU(H₂O)±Δ.

27. A method for reducing line artifacts in a CT image D₁, acquired by scanning an examination subject by rotating at least a focus of an X-ray source around the examination subject and detecting radiation, attenuated by said examination subject, with a detector system which has a plurality of proper detector channels and at least one defective detector channel, and wherein said CT image D₁is subjected to an interpolated filtering for preventing or correcting faulty values of picture elements represented on a circle K₁, said faulty values being caused by the defective detector channel, comprising the steps of:

(a) subjecting picture elements of the CT image D₁to a median filtering, orthogonal to tangents to said circle K₁, and extending through the respective picture element, for producing an image M₁;

(c) producing two resulting images G¹ ₁and G² ₁by filtering picture elements of the difference value image F₁in respective directions of tangents t¹ ₁and t² ₁to the circle K₁extending through the respective picture element, whereby the filtering is conducted along t¹ ₁for the resulting image G¹ ₁and is conducted along t² ¹for the resulting image G² ₁; and

28. A method for reducing line artifacts in a correction image D₂as claimed in claim 27, wherein the number of the defective detector channels equals M (M>1) and wherein i is a running index with i=2 . . . M, further comprising:

(e) subjecting picture elements of the CT image D₁to a median filtering, orthogonal to a straight line extending through the respective picture element and a center of the circle K_i, for producing an image M₁;

(f) forming a difference value image F_i=D_i−M_iby subtracting the image M₁from the CT image D_i;

(g) producing two resulting images G¹ _iand G² _iby filtering picture elements of the difference value image F₁in respective directions of tangents t¹ _iand t² _ito the circle K_iextending through the respective picture element, whereby the filtering is conducted along t¹ _ifor the resulting image G¹ _iand is conducted along t² _ifor the resulting image G² _i; and

(h) forming a correction image D_i+1=D_i−G¹ _i−G² _i, 1≦i≦M, by subtracting the resulting images G¹ _iand G² _ifrom D_i.

29. A method as claimed in claim 27 comprising adjusting a width of said median filtering.

30. A method as claimed in claim 27 wherein the step of filtering said picture elements to produce said resulting images G¹ ₁and G² ₁comprises conducting a running averaging along the respective tangents t¹ ₁and t² ₂.

31. A method as claimed in claim 27 wherein the step of filtering said picture elements to produce the resulting images G¹ ₁and G² ₁comprises employing a sum operator along said tangents t¹ ₁and t² ₁.

32. A method as claimed in claim 27 wherein the step of filtering said picture elements to produce the resulting images G¹ ₁and G² ₁comprises conducting a running averaging and employing a sum operator along said tangents t¹ ₁and t² ₁.

33. A method as claimed in claim 27 comprising limiting values of the picture elements of the CT image D₁to a range 1,000 HU(H₂O)±Δ.

34. A computed tomography apparatus comprising:

a computer connected to said detector system wherein said CT image D₁is subjected to an interpolated filtering for preventing or correcting faulty values of picture elements represented on a circle K₁, said faulty values being caused by the defective detector channel, for reducing line artifacts in said CT image D₁, said computer subjecting picture elements of the CT image D₁to a median filtering, orthogonal to tangents of said circle K₁extending through the respective picture element, for producing an image M₁, and forming a difference value image F₁=D₁−M₁by subtracting the image M₁from the CT image D₁, and producing two resulting images G¹ ₁and G² _l, by filtering picture elements of the difference value image F₁in respective directions of tangents t¹ ₁and t² ₁to the circle K₁extending through the respective picture element, whereby the filtering is conducted along t¹ ₁for the resulting image G¹ ₁and is conducted along t² ₁for the resulting image G² ₁, and forming a correction image D₂=D₁−G¹ ₁−G² ₁by subtracting the resulting images G¹ ₁and G² ₁from D₁.

35. A computed tomography apparatus as claimed in claim 34, wherein the number of the defective detector channels equals M (M>1) and wherein i is a running index with i=2 . . . M, said computer subjecting picture elements of the CT image D₁to a median filtering, orthogonal to a straight line extending through the respective picture element and a center of the circle K_i, for producing an image M₁, and forming a difference value image F_i=D_i−M_iby subtracting the image M_ifrom the CT image D_i, and producing two resulting images G¹ _iand G² _iby filtering picture elements of the difference value image F_iin respective directions of tangents t¹ _iand t² _ito the circle K_iextending through the respective picture element, whereby the filtering is conducted along t¹ _ifor the resulting image Gⁱ ₁and is conducted along t² _ifor the resulting image G² _i, and forming a correction image D_i+1=D_i−G¹ _i−G² _iby subtracting the resulting images G¹ _iand G² _ifrom D_i.

36. A computed tomography apparatus as claimed in claim 34 wherein said computer adjusts a width of said median filtering.

37. A computed tomography apparatus as claimed in claim 34 wherein said computer filters said picture elements to produce said resulting images G¹ ₁and G² ₁by conducting a running averaging along the respective tangents t¹ ₁and t² ₂.

38. A computed tomography apparatus as claimed in claim 34 wherein said computer filters said picture elements to produce the resulting images G¹ ₁and G² ₁by employing a sum operator along said tangents t¹ ₁and t² ₁.

39. A computed tomography apparatus as claimed in claim 34 wherein said computer filters said picture elements to produce the resulting images G¹ _iand G² ₁by conducting a running averaging and employing a sum operator along said tangents t¹ ₁and t² ₁.

40. A computed tomography apparatus as claimed in claim 34 wherein said computer limits values of the picture elements of the CT image D₁to a range 1,000 HU(H₂O)±Δ.