[go: up one dir, main page]

WO2010004781A1 - Dispositif de détection d'ombre anormale, procédé de détection d'ombre anormale et programme - Google Patents

Dispositif de détection d'ombre anormale, procédé de détection d'ombre anormale et programme Download PDF

Info

Publication number
WO2010004781A1
WO2010004781A1 PCT/JP2009/054488 JP2009054488W WO2010004781A1 WO 2010004781 A1 WO2010004781 A1 WO 2010004781A1 JP 2009054488 W JP2009054488 W JP 2009054488W WO 2010004781 A1 WO2010004781 A1 WO 2010004781A1
Authority
WO
WIPO (PCT)
Prior art keywords
region
abnormal shadow
breast
area
density
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/JP2009/054488
Other languages
English (en)
Japanese (ja)
Inventor
剛 小林
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Konica Minolta Medical and Graphic Inc
Original Assignee
Konica Minolta Medical and Graphic Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Konica Minolta Medical and Graphic Inc filed Critical Konica Minolta Medical and Graphic Inc
Priority to JP2010519667A priority Critical patent/JPWO2010004781A1/ja
Publication of WO2010004781A1 publication Critical patent/WO2010004781A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
    • A61B6/50—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications
    • A61B6/502—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment specially adapted for specific body parts; specially adapted for specific clinical applications for diagnosis of breast, i.e. mammography
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30068—Mammography; Breast

Definitions

  • the present invention relates to an abnormal shadow detection device, an abnormal shadow detection method, and a program.
  • CAD Computed-Aided Diagnosis
  • the shadow of a lesion often has a characteristic density distribution, and the CAD detects an image area estimated as a lesion based on such density characteristics as an abnormal shadow candidate area.
  • tumors and microcalcification clusters can be mentioned as characteristic features of the cancerous part of breast cancer, but on a medical image (mammography) taken of the breast, the shadow of the tumor is whitish and round with a density change close to a Gaussian distribution. Appears as a shadow.
  • a microcalcification cluster is a collection of microcalcifications (clustered), and appears on the mammography as a whitish round shadow having a substantially conical density change.
  • the great pectoral muscle taken with the breast is mainly composed of muscle (muscle tissue).
  • the breast is composed of a mammary gland and fat.
  • the mammary gland changes to fat due to the aging of the subject.
  • the mammary gland is photographed white and the fat is photographed black. That is, when the subject is young, the whole breast region is photographed white (dense breast), and when the subject is a middle age, the whole breast region is photographed black (active rest). It will be.
  • an abnormal shadow using an iris filter As described in Patent Document 1 and Patent Document 2, generally, it reacts strongly to a region that is round and has a lower concentration than the surroundings, such as a tumor shadow. That is, when a mammary gland with a high degree of circularity and a thickness appears on the image, the mammary gland of a normal tissue may be erroneously detected as an abnormal shadow.
  • a method of detecting an abnormal shadow candidate based on a curvature obtained from a curved surface indicating a density distribution of a medical image such as mammography has been studied. Abnormal shadows that appear in mammography have more X-ray absorption than normal tissues such as mammary glands and fat (that is, they are photographed white). Candidates for abnormal shadows are detected using the density difference of surrounding mammary glands as an index.
  • the abnormal shadow candidate detection method using an iris filter or the like cannot individually cope with the tissue change of the subject and cannot prevent erroneous detection.
  • the present invention has been made in view of the above problems, and an object of the present invention is to suppress the influence of tissue changes caused by aging of a subject and reduce the occurrence of false detection of abnormal shadows. .
  • an abnormal shadow detection apparatus comprises: Breast area extraction means for extracting a breast area in a breast image; A region determination unit that extracts a pectoral muscle region in the breast region extracted by the breast region extraction unit and determines a region of the extracted pectoral muscle region that refers to a concentration; An abnormal shadow candidate detecting means for detecting a candidate area of an abnormal shadow based on the density of the breast area extracted by the breast area extracting means; A density difference between the average density of the candidate area detected by the abnormal shadow candidate area detection means and the average density of the area determined by the area determination means is calculated, and the candidate area is calculated based on the calculated density difference. Determining means for determining whether or not an abnormal shadow, Is provided.
  • the abnormal shadow detection device is: A storage means for storing a reference value for determining whether or not the candidate area detected by the abnormal shadow candidate area detection means is an abnormal shadow;
  • the determination unit preferably determines that the candidate area is an abnormal shadow when the absolute value of the density difference is equal to or greater than a reference value stored in the storage unit.
  • the abnormal shadow candidate detection unit detects an abnormal shadow candidate region based on a curvature obtained from a density distribution of neighboring pixels within a predetermined region from an arbitrary pixel of interest in the breast region extracted by the breast region extraction unit. It is preferable.
  • the abnormal shadow candidate detection means detects an abnormal shadow candidate area based on a contrast obtained from a density difference between neighboring pixels within a predetermined area from an arbitrary target pixel in the breast area extracted by the breast area extraction means. It is preferable.
  • the abnormal shadow candidate detection means includes a density of a neighboring pixel, a density distribution shape, a density distribution size, and a density distribution area within a predetermined area from any target pixel in the breast area extracted by the breast area extraction means. It is preferable to detect an abnormal shadow candidate region based on any one or a plurality of features of the edges.
  • the region determination unit extracts a pectoral muscle region in the breast region extracted by the breast region extraction unit, and among the extracted pectoral muscle region, a region and / or a lesion that has been imaged by overlapping the breast and pectoral muscles It is preferable to determine a region for which density is referred to by excluding a part region.
  • the abnormal shadow detection method comprises: A breast region extraction step of extracting a breast region in a breast image; A region determination step of extracting a pectoral muscle region in the breast region extracted by the breast region extraction step, and determining a region of the extracted pectoral muscle region that refers to a concentration; An abnormal shadow candidate detection step of detecting a candidate region of an abnormal shadow based on the density of the breast region extracted by the breast region extraction step; A density difference between the average density of the candidate area detected by the abnormal shadow candidate area detection step and the average density of the area determined by the area determination step is calculated, and the candidate area is calculated based on the calculated density difference A determination step of determining whether or not the image is an abnormal shadow, Have
  • the determination step determines that the candidate area is an abnormal shadow when the absolute value of the density difference is equal to or greater than a reference value stored in a storage unit.
  • the abnormal shadow candidate detection step detects an abnormal shadow candidate region based on a curvature obtained from a density distribution of neighboring pixels within a predetermined region from an arbitrary pixel of interest in the breast region extracted by the breast region extraction step. It is preferable.
  • the abnormal shadow candidate detecting step detects an abnormal shadow candidate region based on a contrast obtained from a density difference between neighboring pixels within a predetermined region from an arbitrary pixel of interest in the breast region extracted by the breast region extracting step. It is preferable.
  • the density of neighboring pixels, the shape of the density distribution, the size of the density distribution, the region of the density distribution in the predetermined region range from any target pixel in the breast region extracted by the breast region extraction step It is preferable to detect an abnormal shadow candidate region based on any one or a plurality of features of the edges.
  • the region determination step extracts a pectoral muscle region in the breast region extracted by the breast region extraction step, and a region and / or a lesion in which the breast and pectoral muscles are imaged in the extracted pectoral muscle region. It is preferable to determine a region for which density is referred to by excluding a part region.
  • the program is Computer A breast region extraction means for extracting a breast region in a breast image; A region determination unit that extracts a pectoral muscle region in the breast region extracted by the breast region extraction unit, and determines a region whose density is referred to from the extracted pectoral muscle region; Abnormal shadow candidate detection means for detecting a candidate area for abnormal shadow based on the density of the breast area extracted by the breast area extraction means; A density difference between the average density of the candidate area detected by the abnormal shadow candidate area detection means and the average density of the area determined by the area determination means is calculated, and the candidate area is calculated based on the calculated density difference. Means for judging whether or not an abnormal shadow, To function as.
  • the program is The computer,
  • the determination means further functions as a storage means for storing a reference value for determining whether or not the candidate area detected by the abnormal shadow candidate area detection means is an abnormal shadow,
  • the determination unit preferably determines that the candidate area is an abnormal shadow when the absolute value of the density difference is equal to or greater than a reference value stored in the storage unit.
  • the abnormal shadow candidate detection unit detects an abnormal shadow candidate region based on a curvature obtained from a density distribution of neighboring pixels within a predetermined region from an arbitrary pixel of interest in the breast region extracted by the breast region extraction unit. It is preferable.
  • the abnormal shadow candidate detection means detects an abnormal shadow candidate area based on a contrast obtained from a density difference between neighboring pixels within a predetermined area from an arbitrary target pixel in the breast area extracted by the breast area extraction means. It is preferable.
  • the abnormal shadow candidate detection means includes a density of a neighboring pixel, a density distribution shape, a density distribution size, and a density distribution area within a predetermined area from any target pixel in the breast area extracted by the breast area extraction means. It is preferable to detect an abnormal shadow candidate region based on any one or a plurality of features of the edges.
  • the region determination unit extracts a pectoral muscle region in the breast region extracted by the breast region extraction unit, and among the extracted pectoral muscle region, a region and / or a lesion that has been imaged by overlapping the breast and pectoral muscles It is preferable to determine a region for which density is referred to by excluding a part region.
  • the present invention it is possible to suppress the influence of the tissue change caused by the aging of the subject and reduce the occurrence of erroneous detection of abnormal shadows.
  • FIG. 1 It is a figure which shows the abnormal shadow detection apparatus in this embodiment. It is a flowchart which shows the abnormal shadow detection process which the abnormal shadow detection apparatus shown in FIG. 1 performs. It is a schematic diagram which shows the relationship between a breast area
  • FIG. 1 shows a functional configuration example of the abnormal shadow detection apparatus 10 according to the present embodiment.
  • the abnormal shadow detection apparatus 10 includes a CPU (Central Processing Unit) 11, an I / F (InterFace) 12, an operation unit 13, a display unit 14, a communication unit 15, a RAM (Random Access Memory) 16, A ROM (Read Only Memory) 17, a printer 18, and the like are provided, and each unit is connected by a bus 19.
  • the CPU 11 reads out system programs and various processing programs stored in the ROM 17 and expands them in the RAM 16, and executes various processes including an abnormal shadow detection process described later in cooperation with the expanded programs.
  • the operation of each part of the shadow detection apparatus 10 is centrally controlled.
  • the I / F 12 is an interface for connecting to the image generation apparatus G, and inputs image data generated in the image generation apparatus G to the abnormal shadow detection apparatus 10.
  • the image generation apparatus G is an apparatus that captures a patient's breast as a subject and digitally converts the captured image to generate breast image data.
  • the image generation device G for example, a CR (Computed Radiography) device, an FPD (Flat Panel Detector) device, or the like is applicable.
  • the image generation device G generates breast image data D for one patient and inputs it to the abnormal shadow detection device 10.
  • the operation unit 13 includes a keyboard including cursor keys, numeric keys, and various function keys, and outputs an operation signal corresponding to the pressed key to the CPU 11. Note that a pointing device such as a mouse or a touch panel may be included as necessary.
  • the display unit 14 includes a monitor such as an LCD (Liquid Crystal Display) or a CRT (Cathode Ray Tube), and displays a breast image or the like in accordance with an instruction of a display signal input from the CPU 11.
  • a monitor such as an LCD (Liquid Crystal Display) or a CRT (Cathode Ray Tube)
  • LCD Liquid Crystal Display
  • CRT Cathode Ray Tube
  • the communication unit 15 includes a communication interface such as a network interface card, a modem, and a terminal adapter, and transmits / receives various information to / from external devices on the communication network.
  • the image data may be received from the image generation device G via the communication unit 15 or may be connected to an image server in a hospital via the communication unit 15.
  • the RAM 16 forms a work area for temporarily storing various programs executed by the CPU 11 and data processed by these programs.
  • the ROM 17 functions as a storage unit, and stores various programs executed by the CPU 11 and data such as parameters necessary for execution of processing by the programs or processing results. These various programs are stored in the form of readable program codes, and the CPU 11 sequentially executes operations according to the program codes.
  • Other computer-readable media other than the ROM 17 include a non-volatile memory such as a flash memory such as an SD (Secure Digital) card or a USB (Universal Serial Bus) memory, and a portable recording medium such as a CD-ROM. It is possible to apply. It is also possible to provide various data such as program data according to the present invention via a communication line by superimposing them on a carrier wave.
  • the printer 18 forms and outputs an image on a recording medium such as a film based on the image data under the control of the CPU 11.
  • the abnormal shadow detection apparatus 10 when the breast image data D is input from the image generation apparatus G, an abnormal shadow detection process described below is executed. Note that in the normal case, the image generation apparatus G captures the left and right breasts of the subject. That is, left and right breast image data D are input from the image generation device G, respectively. In the abnormal shadow detection process described below, the abnormal shadow detection process similar to that for one breast is performed for the other breast. Therefore, for convenience of explanation, only the abnormal shadow detection process for one breast image data D will be described. .
  • FIG. 2 shows a flowchart of the abnormal shadow detection process executed in the abnormal shadow detection apparatus 10. This process is realized by a software process in cooperation with the CPU 11 and a program stored in the ROM 17. Note that the pixel value of the breast image data D in the present embodiment indicates a density value. The processing described below is executed by the CPU 11.
  • FIG. 3 schematically shows the breast image data D.
  • step S1 from the breast image data D, an area where X-rays have passed through the subject (hereinafter referred to as breast area Sa) and an area where X-rays did not pass through the subject (hereinafter referred to as outside breast area Sb). Processing to sort is performed.
  • the CPU 11 functions as a breast region extracting unit.
  • a known method may be used for the breast region extraction processing performed in step S1. For example, it is extracted as in the following (1-1) to (1-3).
  • the breast image data D is 12-bit grayscale data
  • each pixel is represented by a pixel value of 0 to 4095.
  • the breast region of the breast image data D is photographed white.
  • the pixel values of the breast region Sa are concentrated between low regions (for example, 0 to 1000).
  • the area other than the subject that is, the region where the X-rays do not pass through the subject are photographed in black.
  • the pixel values of the extramammary region Sb are concentrated between high regions (eg, 4000 to 4095).
  • the breast region Sa and the extramammary region Sb are discriminated based on the tendency to binarize the low region and the high region.
  • each of the fixed sections is simply referred to as a segment).
  • the average pixel value of each segment is compared with a threshold value.
  • This threshold value is a threshold value for considering the breast region Sa, and is stored in the ROM 17 in advance.
  • Each segment is binarized by being compared with a threshold value. That is, processing is performed so that a segment indicating a pixel value smaller than the threshold is “1” and a segment indicating a pixel value larger than the threshold is “0”. Then, the binarized value of each segment is corrected based on the morphology.
  • the breast image data D can be binarized for each segment. Note that binarization may be performed for all pixels without dividing the breast image data D into segments, but in this embodiment, binarization is performed for each segment as described above in order to reduce the processing load. I will do it.
  • the breast image data D is equally divided into two in the vertical direction. That is, the breast image data D is divided into left and right parts. Among the two divided areas, a bright area (that is, an area with many white pixels) is determined. As a determination method, for example, an area having a larger number of segments having a binarized value “1” of the two left and right areas is determined as a bright area. That is, it is determined that the breast is present in the region with the larger number of white pixels.
  • a candidate for the breast region Sa is extracted from the bright regions of the breast image data D. Specifically, first, the size of the area where the binarized value of each segment is “1” is determined. If it is determined that this region is sufficiently wider than a predetermined threshold, the region is set as a candidate for the breast region Da. When it is determined that it is not wider than a predetermined threshold value, the following adjustment process is performed. Four segments (up / down / left / right segments) surrounding a certain segment (referred to as a target segment) are defined as segments of interest.
  • the segment of interest has a bipolar value. Of these, “1” is set. On the contrary, if the average pixel value of the target segment is larger than a predetermined threshold value, “0” of the bipolar values is set in the segment of interest.
  • the adjustment process is performed in this manner, and it is determined again whether the area of the binarized value “1” is sufficiently wide. If a candidate region for the breast region Sa is still not found, the other region (the other of the regions obtained by dividing the breast image data D into two regions, that is, the region that is not bright) is searched. If no candidate is still found, the whole breast image data D is searched. The candidate extracted in this way is defined as a breast region Sa.
  • the pectoral muscle region M1 is extracted from the breast region Sa calculated in step S1 (step S2).
  • a known method may be used for the extraction processing of the pectoral muscle region M1 performed in step S2.
  • the pectoral muscle region M1 extraction process performed in step S2 is extracted as in the following (2-1) to (2-5).
  • the position of each pixel in the breast image data D1 is represented by coordinates (X, Y) with the left-right direction of the breast in the breast image data D1 as the X axis and the vertical direction as the Y axis.
  • the pixel value of the coordinates (X, Y) in the breast image data D1 is represented as V (X, Y).
  • the coordinates of the image end in the X-axis direction are represented as X max
  • the image end in the Y-axis direction is represented as Y max .
  • a skin line SL that is a boundary point between the breast region Sa and the extramammary region Sb in the breast image data D1 is extracted.
  • the boundary between the breast region Sa and the extramammary region Sb calculated in step S1 may be the skin line SL.
  • the skin line SL is as follows. To extract. For each Y coordinate (0 to Y max ) of the breast image data D1, a search is performed in the X-axis direction, and a coordinate S (Y) that maximizes V (X, Y) is extracted. Thereby, the edge in each Y coordinate of the breast image data D1 is extracted. As shown in FIG. 4A, the extracted edge is a boundary point between the breast region Sa and the extramammary region Sb in the breast image data D1, and constitutes a skin line SL.
  • Search lines la0 to la30 having a length of 5 are set.
  • the average value of the pixel values on the search lines la0 to la30 is calculated.
  • the reference point of the search line having the maximum calculated average value is determined as the pectoral muscle line search start point B.
  • the pectoral muscle line search start point B is close to the right end (X max ) of the image, for example, from the right end to 10 pixels, it is determined that there is no pectoral muscle region M1.
  • the calculated maximum average value is smaller than a predetermined threshold value, for example, 300, it is determined that there is no pectoral muscle region M1.
  • a predetermined threshold value for example, 300
  • it is determined that there is no pectoral muscle region M1 for example, an error message or the like is displayed on the display unit 14, and the abnormal shadow detection process ends.
  • FIG. 4B shows an enlarged view near the pectoral muscle line search start point B.
  • the width of the breast image data D in the Y-axis direction is 1/5 in increments of 1 °.
  • Search lines lb0 to lb18 having a length are set. Next, the average value of the pixel values on the search lines lb0 to lb18 is calculated.
  • the search line lbn (n is an integer from 0 to 18) having the maximum calculated average value is determined as the pectoral muscle line L.
  • similar processing is performed using a point that is 1/10 of the width in the Y-axis direction from the pectoral muscle line search start point B as a base point.
  • the same processing is performed with a point that is 1/10 of the width in the Y-axis direction from the point set as the previous base point.
  • the pectoral muscle line L is extracted.
  • a predetermined threshold for example, 300
  • a region where density is not measured (hereinafter referred to as a density reference exclusion region) is determined and excluded from the pectoral muscle region M1 extracted in step S2 (step S3).
  • the pectoral muscle region M1 is extracted from the breast region Sa in step S2, and a region in the pectoral muscle region M1 near the pectoral muscle line L (hereinafter referred to as a mammary gland overlap region T1) is a pectoral muscle region. Not taken into account as the concentration of M1. This is because the mammary gland overlap region T1 may be a region in which the mammary gland and the pectoral muscle are imaged due to poor positioning during imaging.
  • lymphoproliferative region T2 a region in the vicinity of this lesion (hereinafter referred to as lymphoproliferative region T2) is excluded from the density calculation. This is because the concentration of the lymph hypertrophy region T2 and the concentration of tissues such as normal lymph are different.
  • the aforementioned mammary gland overlap region T1 and lymphoproliferative region T2 are determined as the concentration reference exclusion region.
  • the concentration reference exclusion region may be a region that is not regarded as a normal muscle tissue, and is not limited to that in the present embodiment.
  • the ROM 17 may hold a lesion area other than lympho-hypertrophy and a density reference exclusion area preset by the user.
  • step S3 The confirmation and exclusion processing of the mammary gland overlap region T1 and the lymphoproliferative region T2 performed in step S3 is calculated as follows (3-1) to (3-3).
  • the coordinates of the pixel on the pectoral muscle line L are (X max ⁇ X muscle [k], k).
  • k is an arbitrary integer from 0 to Y max .
  • the coordinate of the pectoral muscle line L with the value of Y being 0 (hereinafter referred to as coordinate P1) is (X max -X muscle [0], 0).
  • the value of k when X muscle [k] is 0 is assumed to be Y muscleMAX . That is, in the pectoral muscle line L, the coordinates of the X value of X max (hereinafter referred to as coordinates P2) are (X max , Y muscleMAX ).
  • the coordinate P2 and the X coordinate are the same, and the value of the Y coordinate is 2/3 of the Y coordinate of P2 (hereinafter referred to as the coordinate P3).
  • An area delimited by a straight line connecting P1 (shown by a broken line in FIG. 5 and hereinafter referred to as a straight line L1) is used. That is, in the pectoral muscle region M1, a region having the same X coordinate value and a larger Y coordinate value than the straight line L1 is defined as a mammary gland overlap region T1.
  • the method for determining the mammary gland overlap region T1 is not limited to this as long as it is determined from the pectoral muscle region M1, the pectoral muscle line L, and the like.
  • the pectoral muscle region M2 excluding the mammary gland overlap region T1 from the pectoral muscle region M1 is extracted as a region surrounded by the following expression.
  • lymphatic hypertrophy region T2 included in the pectoral muscle region M2
  • a known method may be used for determining and excluding the lymph hypertrophy region T2.
  • a method of determining the lymph hypertrophy region T2 by performing a determination using a curvature as described below may be used.
  • the curvature is composed of signal components in three directions (three axes of X, Y, and Z) of the coordinates (X, Y) of the pixel included in the pectoral muscle region M2 and the pixel value of the pixel, that is, the density (Z). It is calculated by approximating the normal section of the pixel of interest with a circle from the curved surface obtained from the density distribution and obtaining the radius of the circle.
  • the curvature is an index indicating whether the curved surface is convex or concave. That is, the larger the curvature in the positive direction, the more concave the curved surface, and the larger the curvature value in the negative direction, the convex shape. Therefore, the larger the absolute value of the curvature, the greater the density gradient in the vicinity of the target pixel.
  • Abnormal shadows such as lymphatic hypertrophy are generally classified into concave shapes.
  • the pectoral muscle region M2 is divided into predetermined small regions, and an average value of curvature, a maximum value of curvature, a minimum value of curvature, and the like are calculated as feature amounts for each small region. This feature amount is compared with a preset threshold value.
  • a small region having a feature amount equal to or greater than a threshold value, that is, a large concave region is calculated as a candidate region for lymphoproliferation.
  • the candidate region for lymph hypertrophy is approximated to a circle and the diameter is 10 mm or less, it is recognized as normal lymph.
  • the candidate region is determined as the lymphoproliferative region T2.
  • the lymphoproliferative region T2 calculated as described above is excluded from the pectoral muscle region M2 (hereinafter, this region is referred to as a concentration reference region M3).
  • FIG. 5B shows a lymphoproliferative region T2 and a concentration reference region M3.
  • a region obtained by excluding the lymph hypertrophy region T2 from the pectoral muscle region M2 is the concentration reference region M3.
  • the CPU 11 functions as an area determination unit by the processing in step S2 and step S3.
  • the density reference area M3 may be determined based on the pectoral muscle area T1, and the process of excluding the density reference exclusion area from the density reference area M3 in step S3 is not essential.
  • step S4 the average density D MuscleAve of the density reference region M3 calculated in step S3 is calculated (step S4). Specifically, the pixels included in the density reference area M3 are counted, and the number N of pixels in the density reference area M3 is acquired. A pixel value (density) of an arbitrary pixel included in the density reference region M3 is defined as D (x, y). The average density D MuscleAve is calculated by dividing the sum of the pixel values of the pixels included in the density reference area M3 by the number N of pixels included in the density reference area M3 as in the following equation.
  • a candidate for an abnormal shadow in the breast region Sa is detected (step S5).
  • abnormal shadow candidates are obtained based on the curvature of the density of the pixels included in the breast region Sa. Calculated.
  • step S5 another method may be used as the method for detecting abnormal shadow candidates in step S5.
  • the method for detecting abnormal shadow candidates in step S5 may be used.
  • Japanese Patent Application Laid-Open No. 10-91758 that detects an abnormal shadow from the density distribution shape and the like
  • Japanese Patent Application Laid-Open No. 09-508815 that detects an abnormal shadow based on a contrast difference in density. The method may be used.
  • the iris filter process which is an effective technique for detecting a mass shadow that is one of the characteristic forms particularly in breast cancer, is performed in step S5 of the present embodiment. May be applied.
  • the mass shadow tends to have a slightly lower density value than the surrounding image portion. Therefore, the density value distribution decreases in the density value from the substantially circular periphery toward the center. It has a gradient of density value.
  • the gradient line is concentrated toward the center of the tumor mass. That is, abnormal shadow candidates can be detected based on the shape and size of the density distribution and the features of the edges of the density distribution area.
  • the iris filter calculates the gradient of the image signal typified by this density value as a gradient vector and outputs the degree of concentration of the gradient vector. As a result, a region having a density gradient close to a circle is detected as an abnormal shadow candidate. In step S5, an abnormal shadow may be detected based on the concentration degree of the gradient vector calculated by the iris filter process.
  • an abnormal shadow may be detected using a difference in density contrast.
  • multiple gray level threshold processing is performed on the breast region Sa, accurate region increase and feature analysis are performed to increase specificity, and an abnormal shadow candidate is detected based on the emission angle of the pixel of interest.
  • An abnormal shadow may be detected by determining whether the periphery of the tumor edge, the inside of the tumor, or the periphery of the tumor based on the contrast difference in density calculated by the cumulative edge gradient orientation histogram analysis.
  • the CPU 11 functions as an abnormal shadow candidate detecting means by the processing in step S5.
  • FIG. 6A schematically shows the abnormal shadow candidate areas detected in step S5.
  • the abnormal shadow candidate detected in step S5 is defined as an abnormal shadow candidate region T3. That is, since the curvature of density of the pixels included in the abnormal shadow candidate region T3 is equal to or greater than a certain threshold value, the abnormal shadow candidate is detected in step S5.
  • step S6 the average density of the abnormal shadow candidate region T3 detected in step S5 is calculated.
  • step S6 the method of calculating the average density in step S6 will be described.
  • the following two methods (6-1) and (6-2) can be used to measure the average density of the abnormal shadow candidate region T3.
  • the curvature of the region near the edge of the abnormal shadow candidate region T3 is close to the threshold value.
  • both of the two methods described below are areas where the possibility of abnormal shadows is relatively low in the vicinity of the edge, and therefore the purpose is to measure the average density by excluding the area near the edge. Yes.
  • FIG. 6B schematically shows a case where the edge of abnormal shadow candidate region T3 is approximated by an ellipse (hereinafter simply referred to as approximate ellipse C1). .
  • the approximate ellipse C1 is specified by the major axis r max , the minor axis r min , and the center point O (c, d).
  • the approximate ellipse C1 may be created using any method, for example, by approximating the coordinates of the pixels constituting the edge of the abnormal shadow candidate region T3 by the least square method.
  • 6C schematically shows a circle C2 having a radius r min centered on the center point O (c, d) of the approximate ellipse C1.
  • the area surrounded by the circle C2 may be measured as the average density of the abnormal shadow candidate area T3.
  • the average pixel value of the pixels surrounded by the following formula is measured. The average pixel value is calculated by dividing the sum of the pixel values D ij of the pixels included in the circle C2 by the number of pixels.
  • FIG. 6D schematically shows an area excluding the area near the edge of the abnormal shadow candidate area T3.
  • the coordinates of the edge are (X marginal , Y marginal )
  • the coordinates of the edge are t (shown as an arbitrary constant from 0 to 1) from the center point O (c, d).
  • the average density in the area surrounded by the doubled coordinates is calculated.
  • the average density of the region surrounded by the coordinates (t (X marginal ⁇ c), t (Y marginal ⁇ d)) may be calculated.
  • This average density is calculated from the average pixel value as in (6-1).
  • the average density of the abnormal shadow candidate area T3 calculated as described above is defined as D AbnormalAve .
  • the average density D AbnormalAve of the abnormal shadow candidate area T3 acquired in step S6 is compared with D MuscleAve , so that it is finally determined whether or not the abnormal shadow candidate area T3 is an abnormal shadow.
  • Step S7 the determination is made based on whether or not the absolute value of the difference between the average concentrations D AbnormalAve and D MuscleAve is greater than or equal to a predetermined reference value Th pickup . That is, if the following expression is satisfied, the abnormal shadow candidate area T3 is not determined to be an abnormal shadow in step S7.
  • the abnormal shadow candidate area T3 determined to be an abnormal shadow in step S7 may be superimposed on the breast image data D and displayed on the display unit 14.
  • the CPU 11 functions as a determination unit by the processing in step S7, and the reference value Th pickup is a reference value for determining that the abnormal shadow candidate area T3 is an abnormal shadow.
  • the reference value Th pickup used in step S7 is previously stored in the ROM 17 as described above.
  • the reference value Th pickup may be determined in any way. For example, by displaying the following density relationship table on the display unit 14, the reference value Th pickup may be determined by the judgment of the user (that is, an interpreting doctor or the like). It's okay.
  • FIG. 7 shows a concentration relationship table used when determining the reference value Th pickup .
  • the pectoral muscle density band, the mass density band, the high density mass density band, the mammary gland density band, and the fat density band are displayed on the display unit 14.
  • the abnormal shadow candidates detected by the abnormal shadow detection process executed in the past in the abnormal shadow detection apparatus 10 are classified according to the judgment of the user. That is, the abnormal shadow candidates detected by the abnormal shadow detection process may include erroneously detected normal tissues.
  • the user visually recognizes the candidate for an abnormal shadow displayed on the display unit 14, and diagnoses whether the candidate is a lesion such as a tumor or a normal mammary gland.
  • regions of abnormal shadow candidates are divided into tissues (for example, pectoral muscles, tumors, high-density tumors, mammary glands, and fats), and the average density and tissues of the regions are associated with each other and stored in the ROM 17.
  • tissues for example, pectoral muscles, tumors, high-density tumors, mammary glands, and fats
  • the average density and tissues of the regions are associated with each other and stored in the ROM 17.
  • the average density for each tissue stored in the ROM 17 is referred to, and the average density of the abnormal shadow candidates stored in the ROM 17 and the standard are stored for each tissue classified by the user. Deviation is displayed.
  • an area that is detected as an abnormal shadow by the abnormal shadow detection process, but the user visually recognizes the area displayed on the display unit 16 and determines that the area is a mammary gland is represented as a “breast density band”.
  • the “mammary gland concentration band” indicates that the concentration is concentrated in the vicinity of “1675 ⁇ 185”.
  • 1675 indicates an average value of the density of an area detected as an abnormal shadow but determined as a mammary gland by the user
  • “185” indicates a standard deviation. The same applies to other organizations.
  • the user determines the reference value Th pickup based on the density relationship table.
  • the determined reference value Th pickup is stored in the ROM 17 and used in step S7 of the abnormal shadow detection process to be executed next.
  • the pectoral muscle concentration is distributed in the vicinity of “1141”
  • the high-concentration mass concentration band is distributed in the vicinity of “1305”
  • the reference value Th pickup is “1305-1114”. This can be determined by setting “164” that is “.
  • the reference value Th pickup may be calculated based on the tissue in which the abnormal shadow candidate detected by the abnormal shadow detection process is finally determined by the user and the density of the region.
  • the CPU 11 may calculate the reference value Th pickup every time the abnormal shadow detection process is executed based on the equation.
  • the abnormal shadow detection apparatus 10 in the present embodiment by comparing the density of the pectoral muscle region with the concentration of the region detected as the abnormal shadow candidate, whether the region is an abnormal shadow or not. It can be determined whether or not. In other words, based on the density of the pectoral muscle region with little tissue change due to age, it is possible to determine whether the subject's abnormal shadow candidate is an abnormal shadow or an erroneously detected one. Abnormal shadow detection with little influence of tissue change can be performed.
  • the density reference exclusion region is determined from the pectoral muscle region, and the concentration of the concentration reference exclusion region is not calculated as the average concentration of the pectoral muscle region. That is, since the average density of the pectoral muscle area excluding the area that does not show normal density, such as an imaging error due to poor positioning or a lesion, is calculated, the accuracy of abnormal shadow detection can be further improved.
  • step S7 in the present embodiment the difference between the average density of the pectoral muscle region and the average density of the abnormal shadow candidate is compared with a threshold value to determine whether or not the abnormal shadow candidate is an abnormal shadow. What is necessary is just to judge by the average density of an area
  • each device constituting the abnormal shadow detection device 10 can be changed as appropriate without departing from the spirit of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Physics & Mathematics (AREA)
  • Radiology & Medical Imaging (AREA)
  • Molecular Biology (AREA)
  • Animal Behavior & Ethology (AREA)
  • Optics & Photonics (AREA)
  • Pathology (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Biophysics (AREA)
  • Surgery (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Dentistry (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

Selon l'invention, on diminue l'influence d'une variation de tissu due au vieillissement d'un sujet examiné, et on réduit une fausse détection d'une ombre anormale. Une unité centrale (11) de l'appareil de détection d'ombre anormale (10) extrait une région de sein (Sa) dans des données d'image de sein (D) et extrait une région pectorale (M1) dans la région de sein extraite (Sa). L'unité centrale (11) détecte un candidat d'ombre anormale en fonction de la densité de la région de sein (Sa) et détermine si ou non la région candidate est une ombre anormale en se basant sur la différence entre la densité de la région candidate et la densité d'une région de référence de densité (M3) utilisée en tant que région de référence dans la région pectorale (M1).
PCT/JP2009/054488 2008-07-11 2009-03-10 Dispositif de détection d'ombre anormale, procédé de détection d'ombre anormale et programme Ceased WO2010004781A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2010519667A JPWO2010004781A1 (ja) 2008-07-11 2009-03-10 異常陰影検出装置、異常陰影検出方法、及びプログラム

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2008-180972 2008-07-11
JP2008180972 2008-07-11

Publications (1)

Publication Number Publication Date
WO2010004781A1 true WO2010004781A1 (fr) 2010-01-14

Family

ID=41506903

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2009/054488 Ceased WO2010004781A1 (fr) 2008-07-11 2009-03-10 Dispositif de détection d'ombre anormale, procédé de détection d'ombre anormale et programme

Country Status (2)

Country Link
JP (1) JPWO2010004781A1 (fr)
WO (1) WO2010004781A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012072974A3 (fr) * 2010-11-30 2012-07-26 Matakina Technology Ltd Technique d'imagerie et système d'imagerie

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006340835A (ja) * 2005-06-08 2006-12-21 Konica Minolta Medical & Graphic Inc 異常陰影候補の表示方法及び医用画像処理システム

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006340835A (ja) * 2005-06-08 2006-12-21 Konica Minolta Medical & Graphic Inc 異常陰影候補の表示方法及び医用画像処理システム

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012072974A3 (fr) * 2010-11-30 2012-07-26 Matakina Technology Ltd Technique d'imagerie et système d'imagerie

Also Published As

Publication number Publication date
JPWO2010004781A1 (ja) 2011-12-22

Similar Documents

Publication Publication Date Title
US8675933B2 (en) Breast segmentation in radiographic images
JPH1156828A (ja) 異常陰影候補検出方法および装置
US9808217B2 (en) Breast cancer detection system, breast cancer detection method, breast cancer detection program, and computer-readable recording medium having breast cancer detection program recorded thereon
JP4453321B2 (ja) 医用画像処理装置及びプログラム
JP3852125B2 (ja) ２次元多階調画像におけるパタン検出方法及び装置
JP2005253685A (ja) 画像診断支援装置及び画像診断支援プログラム
US20100284579A1 (en) Abnormal shadow candidate detecting method and abnormal shadow candidate detecting apparatus
JP2005080758A (ja) 画像処理装置
JP6296385B2 (ja) 医用画像処理装置、その医用目的領域抽出方法及び医用目的領域抽出処理プログラム
JP2006325640A (ja) 異常陰影候補の表示方法及び医用画像処理システム
US10061979B2 (en) Image processing apparatus and method
JP2006181025A (ja) 異常陰影検出方法および装置並びにプログラム
JP2005102784A (ja) 画像診断支援装置及び画像診断支援方法
US7835555B2 (en) System and method for airway detection
JP4483250B2 (ja) 画像診断支援装置、画像診断支援方法及びプログラム
WO2010004781A1 (fr) Dispositif de détection d'ombre anormale, procédé de détection d'ombre anormale et programme
JP2006230910A (ja) 画像処理装置及び画像処理方法
WO2023032954A1 (fr) Procédé de traitement d'informations, programme et dispositif de diagnostic d'image
JP2004209059A (ja) 異常陰影候補検出装置
JP2006340835A (ja) 異常陰影候補の表示方法及び医用画像処理システム
JP2002133397A (ja) 異常陰影候補検出装置
JP2010017370A (ja) 濃度調整装置、濃度調整方法、及びプログラム
JP2004283188A (ja) 画像診断支援装置、画像診断支援方法、プログラム及び記憶媒体
JP2006230904A (ja) 画像処理装置及び画像処理方法
JP2008073076A (ja) 乳房ｘ線画像処理装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 09794228

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2010519667

Country of ref document: JP

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 09794228

Country of ref document: EP

Kind code of ref document: A1