JP2012139438A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor which can accurately discriminate among tissues in an object region to be drawn as medical images by using information indicating structural features of the tissues.SOLUTION: The image processor 100, utilizing the fact that CT images of bones have regions where CT values are low that indicate cancellous bones, calculates the feature amount indicating structural features of the object region extracted from CT images, determines whether the object region is a bone or calcification by determining whether the calculated feature amount corresponds to predetermined conditions or not and subjects the object region to different processing according to the information on the result of discrimination. The proportion of the regions where CT values are high to the whole object region, difference in average CT value among split regions of the object region, shapes of CT value curves formed by sampling CT values along arbitrary lines set in the object region or the like can be used as the feature amount.

Description

  The present invention relates to image processing for medical images such as CT images, MR images, and US images.

  Conventionally, diagnosis using medical images such as CT (Computed Tomography) images, MR (Magnetic Resonance) images, US (Ultrasound) images, and the like has been performed. A computer-aided diagnosis apparatus called CAD (Computer Aided Diagnosis) that detects abnormal shadows from medical images has also been developed. For example, there is a report that it is effective to measure the presence or absence of arterial calcification for measuring the degree of progression of arteriosclerosis (Non-patent Document 1, p69 “Conclusion” column, etc.). It is desired to be able to extract accurately.

  On the other hand, conventionally, as a technique for discriminating bone images from other tissues and appropriately removing bone images from medical images, for example, a technique described in Patent Document 1 or the like has been proposed. In Patent Document 1, a pixel value histogram of an image after calibration (a process for equalizing the pixel values of an image other than the bone to the average value ave without changing the pixel value of the bone image) is obtained, and the pixel value histogram is obtained. It describes a technique for obtaining an appropriate threshold for discriminating bones and discriminating bone images from other tissue images (paragraph [0026] of Patent Document 1).

Japanese Patent No. 4206044

Atsuko Kurosawa, two others, “Study of arterial calcification observed in chest CT screening for faculty members”, August 2010, Journal of Japanese Society of CT Screening, Vol. 17, No. 2, p64-p70

However, since calcification and bone have similar CT values, it is difficult to distinguish them. In particular, in the above-described method of Patent Document 1, a threshold value for discriminating bones from other tissues is calculated using a pixel value histogram indicating the appearance frequency of each pixel value. Therefore, it was difficult to accurately discriminate bone images.
As described above, there is a problem that it is difficult to accurately identify different tissues having similar density values but different in medical images.

  The present invention has been made in view of the above-described problems, and is an image that can accurately identify a tissue in a target region to be drawn on a medical image using information indicating the structural features of the tissue. An object is to provide a processing apparatus.

  In order to achieve the above-described object, the present invention provides a calculation unit that calculates a feature amount indicating a structural feature of a tissue for a target region extracted from a medical image, and the feature amount calculated by the calculation unit is a predetermined condition. By determining whether or not the target region is a bone or calcification, different processing is performed on the target region according to the determination result by the determination unit. And an image processing unit.

  According to the present invention, it is possible to provide an image processing apparatus capable of accurately identifying a tissue of a target region drawn on a medical image using information indicating a structural feature of the tissue.

The figure which shows the whole structure of the image processing apparatus 100 Flow chart explaining the flow of bone / calcification discrimination process (1) The figure explaining the difference by a bone and calcification about the field extracted by threshold B further from the inside of the object field extracted by threshold A Flow chart explaining the flow of bone / calcification discrimination process (2) Example of region division based on contour point information of target region Flow chart explaining the flow of bone / calcification discrimination process (3) The figure which shows the example of CT value distribution on the core line 51 set to the object area | region Flow chart explaining the flow of bone / calcification discrimination process (4)

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[First Embodiment]
First, the configuration of an image processing system 1 to which the image processing apparatus 100 of the present invention is applied will be described with reference to FIG.

  As shown in FIG. 1, an image processing system 1 includes an image processing apparatus 100 having a display device 107 and an input device 109, an image database 111 connected to the image processing apparatus 100 via a network 110, and a medical image photographing apparatus. 112.

The image processing apparatus 100 is a computer that performs processing such as image generation and image analysis.
As shown in FIG. 1, the image processing apparatus 100 includes an external device such as a CPU (Central Processing Unit) 101, a main memory 102, a storage device 103, a communication interface (communication I / F) 104, a display memory 105, and a mouse 108. Interface (I / F) 106, and each unit is connected via a bus 113.

  The CPU 101 calls a program stored in the main memory 102 or the storage device 103 to the work memory area on the RAM of the main memory 102 and executes the program, drives and controls each unit connected via the bus 113, and the image processing apparatus Various processes performed by 100 are realized.

  Further, in the present invention, the CPU 101 executes a bone / calcification discrimination process (FIG. 2) described later on an image to be processed.

  The main memory 102 includes a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. The ROM permanently holds a computer boot program, a program such as BIOS, data, and the like. The RAM temporarily stores programs, data, and the like loaded from the ROM, the storage device 103, and the like, and includes a work area used by the CPU 101 for performing various processes.

  The storage device 103 is a storage device that reads and writes data to and from an HDD (hard disk drive) and other recording media, and stores a program executed by the CPU 101, data necessary for program execution, an OS (operating system), and the like. . As for the program, a control program corresponding to the OS and an application program are stored. Each of these program codes is read by the CPU 101 as necessary, transferred to the RAM of the main memory 102, and executed as various means.

The communication I / F 104 includes a communication control device, a communication port, and the like, and mediates communication between the image processing apparatus 100 and the network 110. The communication I / F 104 controls communication with the image database 111, another computer, or a medical image photographing apparatus 112 such as an X-ray CT apparatus or an MRI apparatus via the network 110.
The I / F 106 is a port for connecting a peripheral device, and transmits / receives data to / from the peripheral device. For example, a pointing device such as a mouse 108 or a stylus pen may be connected via the I / F 106.

  The display memory 105 is a buffer that temporarily accumulates display data input from the CPU 101. The accumulated display data is output to the display device 107 at a predetermined timing.

  The display device 107 includes a display device such as a liquid crystal panel and a CRT monitor, and a logic circuit for executing display processing in cooperation with the display device, and is connected to the CPU 101 via the display memory 105. The display device 107 displays display data stored in the display memory 105 under the control of the CPU 101.

The input device 109 is an input device such as a keyboard, for example, and outputs various instructions and information input by the operator to the CPU 101. The operator interactively operates the image processing apparatus 100 using external devices such as the display device 107, the input device 109, and the mouse 108.
Note that the display device 107 and the input device 109 may be integrated, for example, like a touch panel display. In this case, the keyboard layout of the input device 109 is displayed on the touch panel display.

  The network 110 includes various communication networks such as a LAN (Local Area Network), a WAN (Wide Area Network), an intranet, and the Internet, and connects the image database 111, a server, other information devices, and the like to the image processing apparatus 100. Mediate.

  The image database 111 accumulates and stores image data photographed by the medical image photographing device 112. In the image processing system 1 shown in FIG. 1, the image database 111 is connected to the image processing apparatus 100 via the network 110, but the image database 111 is provided in, for example, the storage device 103 in the image processing apparatus 100. May be.

Next, the operation of the image processing apparatus 100 according to the first embodiment will be described with reference to FIGS.
The CPU 101 of the image processing apparatus 100 reads a program and data related to the bone / calcification discrimination process shown in FIG. 2 from the main memory 102, and executes a process based on the program and data. Further, it is assumed that the CT image to be processed is read from the image database 111 or the storage device 103 and stored in the main memory 102.

First, the CPU 101 performs threshold processing using the threshold A, and extracts a target area from the CT image (steps S101 and S102).
The threshold A may be given by any method, and may be a predetermined value or an input value. However, since bone or calcification shows a larger CT value than other soft tissue, the threshold A is set to a value that distinguishes between calcification and bone and soft tissue. In the threshold process, an area having a pixel value equal to or greater than the threshold A is extracted.
In the following description, the extracted target area is referred to as AR _i (i = 1, 2,..., N).

Next, the CPU 101 sets a threshold value B higher than the threshold value A (step S103), and designates a target region i = 1 (target region to be processed first) AR ₁ from a plurality of target regions AR _i (step S103). S104), obtained by thresholding the specified target area AR ₁ inside the threshold B, the extraction region BR ₁ (high CT value region) (step S105).
The threshold value B may be given by any method, and may be a predetermined value or an input value.

CPU101 calculates the ratio of the area of the extraction region BR ₁ with respect to the target area AR ₁ (area ratio r) (step S106).

Due to its structure, bone has a region called spongy at the center and a region called dense around it. The sponge quality shows a relatively low CT value, and the dense region shows a higher CT value than the sponge quality. The sponge quality is extracted by the threshold value A described above, but is not extracted by the threshold value B. Therefore, a group of low CT regions that are not extracted by the threshold B exist inside the target region AR _i .
On the other hand, calcification generally has a relatively uniform CT value in the entire region, and does not have a group of low CT value regions inside (low CT values may be scattered).
Therefore, when the ratio (area ratio r = BR _i / AR _i ) of the extraction region BR _i (high CT value region) to the target region AR _i is compared between the calcification and the bone, the bone shows a smaller value.

The CPU 101 compares the area ratio r with a predetermined value (step S107). If the area ratio r is smaller than the predetermined value (step S107; YES), the target area (AR ₁ ) being processed is determined as a bone. , subjected to processing to be applied to the case of bone for the target area AR _1. For example, to remove the target region AR ₁ as long as the process for deleting the bone image from the image, leaving the bone image reversed, leaving the target area AR ₁ as long as the process for deleting the calcifications (step S108).

If it is determined that the area ratio r is equal to or greater than a predetermined value (step S107; NO), the target area (AR ₁ ) being processed is determined to be calcified, and the target area AR _{1 is} calcified. Apply the processing that should be applied to the case. For example, the target area AR ₁ as long as the process for deleting the bone image from the image leaves because it is calcified, deletes the target area AR ₁ as long as the process for deleting the calcifications reversed (step S109).

The CPU 101 determines whether or not all target areas AR _i have been specified (step S110). If there is a target area AR _i that has not been specified yet (step S110; NO), the counter i is set to “1”. addition to, the following target regions AR _2, repeats the processing of steps S105~ step S109.
All target areas AR _i are designated, threshold processing by threshold B (step S105), calculation of area ratio r (step S106), discrimination of bone or calcification (step S107), and an image based on the discrimination result When the processing (step S108, step S109) ends (step S110; YES), a series of bone / calcification determination processing ends.

FIG. 3 is a diagram for explaining the target area AR _i extracted in the bone / calcification discrimination process of FIG. 2 and the CT value distribution inside the target area.
As illustrated in FIG. 3, it is assumed that target areas AR _{1 to} AR ₇ are extracted from the CT image 2 by a threshold A. Among them, it is assumed that FIG. 3 (b) high CT value region BR ₁ as shown in the left figure is extracted when the threshold processing by the threshold value B the internal target region AR ₁ shown in 3 (a) left figure, also , it is assumed that FIGS. 3 (a) When the inside of the target region AR ₂ shown in the right figure thresholding with a threshold value B Fig 3 (b) high CT value region BR ₂ as shown in the right figure are extracted.

Since the inside of the target region AR ₁ is low CT value region which is not extracted in the threshold value B, the area ratio _{r (= BR 1 / AR 1} ) becomes smaller than the predetermined value, the target area AR ₁ is determined to bone. Also, the object area AR _2, the extraction region BR ₂ by the threshold B is almost the same area extracting a target area AR _2, the area ratio r becomes greater than a predetermined value, the target area AR ₂ is calcified and discrimination.
Similarly, for all the target areas AR _i , bone and calcification are determined, and processing according to the determination result is performed.

  In the above-described example, one threshold value B is used as a threshold value for discretely multi-leveling the density value distribution in the target region. However, the threshold value is not limited to this, and two or more threshold values B, C,. May be set, the area ratio of the extraction region based on each threshold value may be obtained, and whether the target region is bone or calcification may be identified based on the size of the area ratio. Moreover, although the example which distinguishes a bone | frame and calcification was shown in the above-mentioned example, this Embodiment can also be applied to identification of another structure | tissue.

In this case, for example, the area ratio r ′ = CR _i / AR _i between the region CR _i (not shown) extracted when the target region AR _i is threshold-processed with the threshold value C greater than the threshold value B is set as the target region AR _i. Then, the size of the area ratio r ′ is further determined. As a result, the interior of the target area is divided into three or more different areas. Since the structure of the tissue is clarified by the ratio of these regions, it is possible to determine which tissue characteristic is shown. As described above, by determining the size of the area ratio of each region extracted by a plurality of threshold values, it is possible to quantitatively determine which tissue features the target region shows, and to accurately determine the tissue. .

As described above, in the first embodiment, the image processing apparatus 100 uses the high CT value region in the entire target region as the feature amount representing the internal CT value distribution for the target region AR _i extracted from the CT image 2. The ratio (area ratio r) of (extraction area BR _i by threshold B) is calculated, and the target area is a bone based on the size (area ratio r) of this high CT value area (extraction area BR _i ). It is determined whether it is calcified or not, and different processing is performed on the target area AR _i according to the determination result.

  Thus, if the CT value distribution inside the target region is expressed by the size (area ratio r) of the high CT value region (or low CT value region) in the entire target region, the structural features inside the target region And the characteristics of the composition can be expressed numerically, and it is possible to accurately identify whether the target region is bone or calcification. Then, it is possible to perform appropriate processing for each identified different tissue, and more accurately perform various processing performed on the original medical image, such as drawing of a medical image, extraction of lesion candidates, and the like. Is possible.

In particular, it is effective for identifying tissues having similar structural features but similar CT values, such as bone and calcification. For example, it is suitable for distinguishing a tissue having two or more different compositions (spongeous, dense) such as bone from a uniform tissue such as calcification. Further, the inside of the target area AR _i is multi-valued by a plurality of thresholds, and the area ratio of each area BR _i , CR _i ,... And AR _i (for example, AR _i : BR _i : CR _i or BR _i). / AR _i and CR _i / AR _i etc.) can be calculated, and it becomes possible to identify tissues having three or more different compositions.

  In the above-described embodiment, a two-dimensional cross-sectional view is assumed as an example of a medical image, but may be applied to a three-dimensional volume image obtained by stacking a series of tomographic images. In this case, the volume ratio may be obtained instead of the area ratio described above.

[Second Embodiment]
Next, a second embodiment of the operation of the image processing apparatus 100 according to the present invention will be described with reference to FIGS.

In general, it is assumed that the bone drawn on the CT image or the like has poor symmetry of the internal CT value distribution.
Therefore, the image processing apparatus 100 according to the second embodiment divides the inside of the target region based on the contour point information, and uses the average density value comparison value (difference value) for each divided region as a feature of tissue discrimination. As a quantity, it is determined whether the target region is bone or calcification.

As shown in FIG. 4, first, the CPU 101 performs threshold processing using the threshold A as in steps S <b> 101 to S <b> 102 of FIG. 2, and extracts a target area AR _i from the CT image (step S <b> 201).

Next, the CPU 101 designates a target area i = 1 (target area to be processed first) AR ₁ from among a plurality of target areas AR _i (step S202), and a plurality of contour points of the designated target area AR ₁ The two farthest contour points P1 and P2 are obtained (step S203).

Furthermore, CPU 101 is a contour point P1, the target area AR ₁ of the two at the midpoint P2 divided regions X, Y (hereinafter, X region, called the Y region) is divided into (step S204), the average CT value of the X region The average CT value avCT_Y of the avCT_X and Y regions is obtained (step S205). Then, the absolute value difference abs (avCT_X−avCT_Y) of the average CT value of each divided region is calculated, and the magnitude of the absolute value difference is determined (step S206).

When the absolute value difference of the average CT values in each region is larger than the predetermined value (step S206; YES), it can be determined that the CT value distribution has low symmetry. That is, the CPU 101 determines that the target area (AR _i ) being processed is a bone, and performs a process to be applied to the target area AR _{i in} the case of a bone. For example, if the process is to delete bone from the image, the target area AR _i is deleted. Conversely, if the process is to delete bone, the target area AR _i is left (step S207).

Further, when it is determined that the absolute value difference of the average CT value of each region is equal to or less than the predetermined value (step S206; NO), the CPU 101 determines that the target region (AR _i ) being processed is calcification, Processing to be applied in the case of calcification is performed on the target area AR _i . For example, if the process is to delete a bone from the image, the target area AR _i is calcified, so it is left. If the process is to delete the calcification, the target area AR _i is deleted (step S208).

The CPU 101 determines whether or not all target areas AR _i have been specified (step S209). If there is a target area AR _i that has not been specified yet (step S209; NO), the counter i is set to “1”. In addition, the processing of step S203 to step S208 is repeatedly executed for the next target region.
All the target areas AR _i are designated, and all the target areas are divided based on the contour point information (step S203, step S204), the average CT value of each divided area X, Y is calculated (step S205), and each divided area X , Y absolute CT difference calculation and determination (step S206), and when the image processing based on the determination result (step S207, step S208) ends (step S209; YES), a series of bone / calcification determination processing Exit.

FIG. 5 is a diagram for explaining region division based on the target area AR _i extracted in the bone / calcification determination process (2) of FIG. 4 and the contour point information of the target area AR _i .
As shown in FIGS. 5A and 5B, the target area AR _i is divided into two in the longitudinal direction by the processing of step S203 to step S204, and one is an X area and the other is a Y area. In the case of calcification, since the entire interior of the target area AR _i shows a uniform density value distribution due to its structural characteristics, the average CT values of the X area and the Y area are almost equal, and the absolute value difference is close to zero. Indicates the value. On the other hand, in the case of bone, since the internal CT value distribution is often asymmetric compared to the case of calcification due to its structural characteristics, a difference is likely to occur in the average CT value between the X region and the Y region. Therefore, if the absolute value difference of the average CT values between the X region and the Y region shows a value larger than a predetermined value, it is determined as a bone.

Note that, in the bone / calcification discrimination process of the second embodiment, the method of dividing the area is arbitrary.
For example, as shown in FIG. 5 (c), even when a straight line connecting the farthest contour points P1 and P2 passes outside the target area AR _i , it is divided by dividing lines passing through the intermediate points of the contour points P1 and P2, What is necessary is just to obtain | require the difference of the average CT value of a division area.
Further, as shown in FIGS. 5D and 5E, even when the target region is circular and there are a plurality of combinations of the farthest contour points, if the division is performed at the midpoint of the straight line connecting the combinations of the contour points. Good.

Furthermore, the target area may be divided into three or more areas.
For example, as shown in FIG. 5 (f), the target area AR _i is divided into three in the longitudinal direction, the average CT value is calculated for each of the areas X, Y, and Z, and the average of a certain reference area (for example, Y area) is calculated. The absolute value difference from the CT value is obtained, and when the difference in average CT value in each region is larger than a predetermined value, the target region AR _i is determined as a bone, and the difference in average CT value in each region Is smaller than the predetermined value, it is determined that the target area AR _i is calcified.

As described above, in the second embodiment, for the target area AR _i extracted from the CT image, the target area AR _i is divided into a plurality of areas as the feature amount representing the internal CT value distribution, and each division is performed. The average CT values in the regions are compared (differed), and based on the comparison result of the average CT values, it is determined whether the target region AR _i is a bone or calcification, and the target region AR _i is determined according to the determination result. Different processing is performed.

  In this way, if the characteristics of the CT value distribution inside the target region are expressed by the difference value between the average CT values of at least two divided regions, the structural asymmetry of the tissue can be represented by the asymmetry of the CT value distribution inside the target region. Since it can be clearly quantified as (that is, position dependency), it is possible to accurately identify whether the target region is bone or calcification.

  In the above-described embodiment, an example in which the average CT value of each divided region is compared has been shown. You may compare CT value distribution of an area | region.

[Third Embodiment]
Next, a third embodiment of the operation of the image processing apparatus 100 according to the present invention will be described with reference to FIGS.

  The image processing apparatus 100 according to the third embodiment creates a CT value curve (4a and 4b in FIG. 7) indicating the relationship between the position on the line and the CT value along an arbitrary line set inside the target region. Based on the shape of the CT value curve, it is determined whether the target region is bone or calcification.

As shown in FIG. 6, first, the CPU 101 performs threshold processing using the threshold A, and extracts a target area AR _i from the CT image (step S301). The extraction process of the target area AR _i is the same as steps S101 to S102 in the first embodiment.

Next, CPU 101 specifies the AR ₁ (the target area processing first) i = 1 of the target area from among a plurality of target areas AR _i (step S302), any line in the specified target area AR ₁ Setting is made (step S303). Line set here, a curve or straight line traversing or crossing the target area AR _1. For example, (one-dot chain line 41 in FIG. 7) core of the target area AR ₁ and, (solid arrows 42 to 46 in FIG. 7) a line perpendicular to the core wire and the like.

  The CPU 101 samples the CT value along the set line (core line 41 in the example of FIG. 7), and obtains a CT value curve indicating the relationship between the position on the core line and the CT value (step S304).

FIG. 7 is an example of the graph 4 showing the relationship between the position on the core line 41 and the CT value for the target areas AR ₁ and AR ₂ , 4 a being a CT value curve when the target area is a bone, and 4 b being It is CT value curve in case an object area | region is calcification.
As shown in FIG. 7, when the target region is a bone, there is a low CT value region indicating cancellous material inside, and thus the CT value curve 4 a has a depression near the center. On the other hand, when the target region is calcification, since there is no low CT value region such as sponge, a relatively uniform CT value distribution is shown as shown in the CT value curve 4b.

The CPU 101 analyzes the shape of the CT value curve obtained in step S304, and determines whether or not the depth of the depression near the center of the CT value curve is greater than a predetermined value. As a result of the determination, if the depth of the depression of the CT value curve is larger than the predetermined value (step S305; YES), the target area (AR _i ) being processed is determined as a bone, and the target area AR _{i is} boned. The processing to be applied in the case of For example, if the process is to delete a bone from the image, the target area AR _i is deleted. Conversely, if the process is to delete the calcification by leaving the bone, the target area AR _i is left (step S306).

Further, when the depth of the depression of the CT value curve is equal to or smaller than the predetermined value (step S305; NO), the target area (AR _i ) being processed is determined as calcification, and lime for the target area AR _i is determined. The processing that should be applied in the case of conversion is applied. For example, if the process is to delete a bone from the image, the target area AR _i is left because it is calcified, and conversely, only the bone is left, and if the process is to delete calcification, the target area AR _i is deleted (step S307). ).

The CPU 101 determines whether or not all target areas AR _i have been designated (step S308). If there is a target area AR _i that has not been designated yet (step S308; NO), the counter i is set to “1”. In addition, the processing of step S303 to step S305 is repeatedly executed for the next target area.
All the target areas AR _i are specified, and a line for traversing or traversing the target area for all the target areas is set (step S303), a CT value curve is created (step S304), and a determination is made based on the depth of the depression of the CT value curve. (Step S305) When the image processing based on the determination result (Step S306, Step S307) ends (Step S308; YES), the series of bone / calcification determination processing ends.

Note that the number of vertical lines (or transverse lines) set in the target region is not limited to one and may be plural. For example, as shown in FIG. 7, a plurality of straight lines 42, 43,... Orthogonal to the core wire 41 are set, and a CT value curve is created by plotting CT values along the respective straight lines 42, 43,. (Not shown). In this case, it is possible to obtain the same number of CT value curves as the number of set straight lines 42, 43,..., All the CT value curves are judged, and the CT value curve having a depression larger than a predetermined depth. Is equal to or greater than a predetermined ratio (for example, 1/3 or more), the target region is determined as a bone.
Although it is actually a bone like a straight line 46 shown in FIG. 7, a low CT value region may not exist in some places as in calcification. Even in such a case, it is possible to accurately discriminate by creating and discriminating a plurality of CT value curves.

As described above, in the third embodiment, the CT value is sampled along the arbitrary line set inside the target region as the feature amount representing the internal CT value distribution for the target region extracted from the CT image. A CT value curve is obtained, it is determined whether the target region is a bone or calcification based on the shape of the CT value curve, and different processing is performed on the target region according to the determination result.
If the characteristics of the CT value distribution inside the target area are represented by a CT value curve in this way, it is possible to clearly represent the position dependence of the CT value distribution caused by the structural characteristics of the target area, and the target area is a bone. Or calcification can be accurately identified.
Further, if a plurality of the above-mentioned arbitrary lines are set and the shape is determined for a plurality of CT value curves, the characteristics of the CT value distribution can be examined in more detail, and bone and calcification can be more accurately determined.

As described above, as shown in the first to third embodiments, the image processing apparatus 100 according to the present invention focuses on the difference in composition (structural characteristics) between bone and calcification and extracts a target region from a CT image. The feature amount representing the internal CT value distribution is calculated with respect to, and it is determined whether or not the target region is a bone or calcification by determining whether or not the calculated feature amount satisfies a predetermined condition. Different processing is performed on the target area according to the result.
Then, as the feature amount, the ratio of the high (or low) CT value region in the entire target region (area ratio of the first embodiment), the average CT value difference of each divided region of the target region (second embodiment) Absolute value difference), the shape of a curve obtained by sampling the CT value along an arbitrary line set in the target region (the presence or absence of a dent in the curve of the third embodiment), and the like.

Further, at least two of the feature amounts shown in the first to third embodiments may be combined to determine whether the target region is bone or calcification.
For example, as illustrated in the flowchart of FIG. 8, the CPU 101 extracts the target area AR _i from the CT image using the threshold A, and executes Steps S <b> 105 to S <b> 107 of FIG. 2 for each extracted target area AR _i . That is, the ratio of the high CT value region in the entire target region (area ratio r of the first embodiment) is calculated, and whether the target region is bone or calcification based on the ratio (area ratio). It discriminate | determines (step S403).

  If the bone is determined (step S403; bone), steps S203 to S206 in FIG. 4 are further executed. That is, the target area is divided based on the outline information of the target area, the average CT value difference (absolute value difference of the second embodiment) of each divided area is calculated, and the average CT value difference (absolute value difference) is calculated. It is determined whether the target region is a bone or calcification based on the size (step S404).

  If it is determined as a bone (step S404; bone), steps S303 to S305 in FIG. 6 are further executed. That is, an arbitrary line that traverses or traverses the target region is set, and the shape of the CT value curve obtained by sampling the CT value along the line (the presence or absence of the depression of the CT value curve of the third embodiment) is obtained, and the shape Based on this, it is determined whether the target region is a bone or calcification (step S405).

If it is determined as a bone (step S405; bone), the target area AR _i being processed is determined as a bone, and a process to be applied to the area is applied to the area. For example, if the process is to delete a bone from the image, the target area AR _i is deleted. Conversely, if the process is to delete the calcification by leaving the bone, the target area AR _i is left (step S406).

Moreover, when it is discriminated as calcification in any of Steps S403 to S404, the target area AR _i being processed is discriminated as calcification, and the process to be applied to the area in the case of calcification is performed. Apply. For example, if the process is to delete a bone from the image, the target area AR _i is calcified, so it is left. Conversely, if the process is to delete the calcification, the target area AR _i is deleted (step S407).

The CPU 101 determines whether or not all target areas AR _i have been specified (step S408). If there is a target area AR _i that has not been specified yet (step S408; NO), the counter i is set to “1”. In addition, the processing of step S403 to step S407 is repeatedly executed for the next target region.
When all the target areas AR _i are designated and the bone / calcification determination and the image processing based on the determination result are completed for all the target areas (step S408; YES), the series of bone / calcification determination processes is ended. .

  In the flowchart of FIG. 8, any one of step S403, step S404, and step S405 may be omitted. Further, the processing order of step S403, step S404, and step S405 may be changed.

  In addition, it is obvious that those skilled in the art can arrive at various changes or modifications within the scope of the technical idea disclosed in the present application, and these naturally belong to the technical scope of the present invention. It is understood.

1. Image processing system 100 Image processing apparatus 101 CPU
102... Main memory 103... Storage device 104... Communication I / F
105... Display memory 106... I / F
107... Display device 108... Mouse 109... Input device 2... CT image AR _i. Target area BR _i extracted by threshold A ... Extracted areas X, Y extracted by threshold B ... Divided areas P1, P2 ... Contour points of target area Of the farthest contour point 32 of the soft tissue 4 ... graph 41 ... core 4a ... CT value curve (For bones)
4b ... CT value curve (in the case of calcification)
42, 43, 44, 45, 46 ... Straight line perpendicular to the core wire

Claims (4)

A calculation means for calculating a feature amount indicating a structural feature of the tissue for a target region extracted from a medical image;
Discriminating means for discriminating whether the target area is bone or calcification by determining whether or not the feature amount calculated by the calculating means meets a predetermined condition;
Image processing means for performing different processing on the target area according to the determination result by the determination means;
An image processing apparatus comprising: The calculating means calculates the feature amount,
Threshold processing means for performing threshold processing on the inside of the target region using one or more thresholds;
A ratio calculating means for calculating a ratio of each extraction area extracted by the threshold processing means to the target area,
The image processing according to claim 1, wherein the determination unit determines whether the target region is a bone or calcification based on the size of the ratio calculated by the ratio calculation unit. apparatus. The calculating means calculates the feature amount,
Dividing means for dividing the target area into a plurality of divided areas based on contour point information of the target area;
Statistical value calculation means for calculating the statistical value of the density value for each divided region divided by the dividing means,
The said discrimination | determination means compares the said statistical value of each division area, and discriminate | determines whether the said object area | region is a bone or calcification based on the comparison result. Image processing device. The calculating means calculates the feature amount,
Line setting means for setting a predetermined line in the target area based on the outline information of the target area;
CT value curve generating means for generating a CT value curve obtained by sampling the pixel value of each pixel on the line along the line set by the line setting means,
The image processing according to claim 1, wherein the determination unit determines whether the target region is a bone or a feature amount calcification indicating a structural feature based on a shape of the CT value curve. apparatus.

Images