JP2001357381A - Method and system for detecting abnormal shade candidate - Google Patents

Abstract

PROBLEM TO BE SOLVED: To permit the detection level of an abnormal shade candidate to be the proper one in accordance with past abnormal shade candidate detecting information in an abnormal shade candidate detecting system. SOLUTION: The system is provided with an abnormal shade candidate detecting means 1 for detecting the abnormal shade candidate based on mammography P. The system is also provided with a past information obtaining means 2 for receiving the input of information concerning the detection of the abnormal shade candidate in the past about a subject and a detection level changing means 3 for changing the detection level of the abnormal shade candidate by the means 1 with respect to a specified area corresponding to a past abnormal shade candidate detecting area which is inputted to the means 2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormal shadow candidate detecting method and an abnormal shadow candidate detecting system, and more particularly, to a detecting method and a detecting system with improved diagnostic performance.

[0002]

2. Description of the Related Art Conventionally, in a medical field, a radiation image of a patient is photographed, and a film on which the obtained radiation image is recorded is read on a shakasten for interpretation.
They are used to diagnose injuries and other conditions. In recent years, instead of directly exposing a radiation image to a film,
The radiation image is stored in a stimulable phosphor sheet (radiation image conversion panel) for storing and storing the radiation image, and the excitation light is applied to the stimulable phosphor sheet to emit the stimulated emission light in a light amount corresponding to the stored and stored radiation image. To obtain a digital image signal, and thereafter, the digital image signal is reproduced as a visible image by printing it on a recording medium such as a film or displaying it on an image display device such as a CRT. (Radiation image recording and reading system (computed radiography);
No. 29, No. 56-11395, No. 56-11397).

According to such a radiation image recording and reading system, by performing various signal processing on the digital image signal, the gradation characteristics and frequency characteristics of the output visible image are more suitable for image reading than before. The diagnostic performance based on images can be improved.

[0004] The obtained digital image signal is analyzed using a computer, and abnormal shadows such as tumor shadows and microcalcification shadows appearing in the image are automatically detected. A system that secures a certain detection level even when the degree is low and is useful for diagnosis has been developed (abnormal shadow candidate detection system; Japanese Patent Application Laid-Open No. 8-294479, etc.).

This system automatically detects a candidate for a tumor shadow in an image by evaluating the degree of concentration of a density (signal value) gradient vector of a digital image signal (mammography) of a breast obtained mainly in a breast cancer screening. Or morphological operations (dilation processing,
By performing an erosion process, an opening process, a closing process, etc.) to automatically detect candidates for microcalcification shadows, and to detect abnormal shadow candidates such as tumor shadows and microcalcifications. An abnormal shadow candidate detected by this system is displayed on a CRT or a liquid crystal display device by, for example, marking it in a rectangular ROI (region of interest) frame on mammography,
By printing on a diagnostic film, it can be used for diagnosis.

[0006]

As described above, the abnormal shadow candidate detection system automatically detects abnormal shadow candidates on the basis of the obtained image, thereby improving the image reading ability of the image interpreter. Regardless, since an abnormal shadow candidate can be detected at a certain detection level, it is a very effective system in suppressing the oversight of the abnormal shadow candidate, but the abnormal shadow candidate is determined based only on the image. Because it is to be detected, for example, information on the detection of an abnormal shadow candidate obtained in the past, that is, when an abnormal shadow candidate is detected in the past, the size and type of the abnormal shadow candidate in the detected part And malignancy, and diagnostic information about the abnormal shadow candidate (including an instruction to follow up and a diagnostic result indicating that the shadow is not a true abnormal shadow), etc. Not reflected at all the detection of the abnormal shadow candidate for the image.

In general, the results of detection of abnormal shadow candidates in the past have some influence on the diagnosis of the current subject, and for example, there is a history that abnormal shadow candidates have been detected as past detection results. In such a case, the abnormal shadow candidate and its trace often remain even now. However, according to the conventional abnormal shadow candidate detection method, the abnormal shadow candidate may not be detected from the image of the current subject.

That is, in the process of detecting an abnormal shadow candidate, a state that the abnormal shadow uniquely shows in the image to be processed is represented as a characteristic value, and the characteristic value is finally subjected to a threshold value processing to obtain an abnormal shadow. Since it is detected as a candidate, for example, it is an abnormal shadow candidate detected in a past image that is not the image, and although it is present in the image, the abnormal shadow candidate becomes smaller during the healing process. For this reason, in the abnormal shadow candidate detection process in the current image, the threshold process may not be performed to detect an abnormal shadow candidate.

On the other hand, even if an image interpreter such as a doctor judges that there is a high possibility that an abnormal shadow is present even in the current subject, based on information on the detection of an abnormal shadow candidate in the past. In the abnormal shadow candidate detection process, an abnormal shadow that has not reached a detection level defined by a preset threshold or the like may not be detected as described above.In this case, since it is not detected as an abnormal shadow candidate, Abnormal shadows may be overlooked as completely cured. Diagnosis that the abnormal shadow candidate detected in the previous examination result (the result of the abnormal shadow candidate detection processing based on the past image) is not high in malignancy but needs to be followed up in the future. In this case, there is a possibility that the follow-up observation instruction itself may be overlooked because it is not detected in the present result.

As described above, information on the detection of past abnormal shadow candidates is a very important factor in determining the current state of a subject.

The present invention has been made in view of the above circumstances, and the reliability of abnormal shadow candidate detection can be improved in consideration of information on past abnormal shadow candidate detection for a subject.
It is an object of the present invention to provide an abnormal shadow candidate detection method and a detection system that are higher than before.

[0012]

SUMMARY OF THE INVENTION An abnormal shadow candidate detecting method and an abnormal shadow candidate detecting system according to the present invention provide an abnormal shadow candidate detected in the past in accordance with information on the detection of a past abnormal shadow candidate for a subject. For the same position (region) as the detection position of, the detection level of the abnormal shadow candidate is increased to make it easier to detect the abnormal shadow candidate at the position, or the abnormal shadow candidate detected in the past is not a true abnormal shadow. If it is already known that the detected tissue is a normal tissue / structure, the detection level of the abnormal shadow candidate is lowered at a position (area) substantially the same as the detection position where the detection was erroneously performed, and the erroneous detection is performed. For example, the reliability of the detection is improved as compared with the related art.

That is, an abnormal shadow candidate detecting method according to the present invention is the abnormal shadow candidate detecting method for detecting an abnormal shadow candidate in an image based on the image of the object. According to another aspect of the present invention, a detection level of at least a part of the image corresponding to the specific part is changed according to information on candidate detection.

Here, the information relating to the detection of a past abnormal shadow candidate is information representing a result of performing a process of detecting an abnormal shadow candidate on the past image based on an image obtained in the past (past image). And various output values in the process of the detection process (for example, the output value of the iris filter process, the output value of the morphology calculation process, the degree of malignancy, etc.) regardless of whether or not the candidate is an abnormal shadow actually detected. , Abnormal shadow candidate (regardless of whether it is finally detected)
Means the detection position (part, region) of the subject, the evaluation value such as the degree of malignancy of the abnormal shadow candidate, the size, the type (such as whether it is a tumor shadow candidate or a microcalcification shadow candidate, and the like).

For example, when the information on the detection of the past abnormal shadow candidate indicates that the abnormal shadow candidate has been detected in the past at the specific portion, at least the abnormal shadow candidate corresponding to the specific portion is detected. Changing the detection level for at least a portion of the image (main image) that is a candidate for the candidate detection processing so as to increase the detection rate of the abnormal shadow candidate has an effect of preventing missed interpretation. On the contrary, in order to exert the effect of suppressing erroneous detection, the information on the detection of the past abnormal shadow candidate is the abnormal shadow candidate previously detected in the specific part. Indicates a diagnostic result that is not a true abnormal shadow, the abnormal shadow in the part of the main image corresponding to the specific part To lower the detection rate of, may be changed to detect the level of the portion. In addition, even in a region that is not finally detected as an abnormal shadow candidate in a past image, the output value of the detection processing for that region has a slight threshold value for determining whether or not to detect the region as an abnormal shadow candidate. If it was not detected as an abnormal shadow candidate because it was less than this, the current image that is the target of abnormal shadow candidate detection processing (this image)
May be changed so as to increase the detection level.

In changing the detection level of the abnormal shadow candidate to a higher level, the same type as the type of the abnormal shadow candidate detected in the previous (past) period (for example, whether it is a tumor shadow candidate or a microcalcification shadow candidate). It is preferable to change the detection level so as to increase only the detection rate of the abnormal shadow candidate.
This is because, if the types of the shadows are different, the abnormal shadow candidates detected in the past and the abnormal shadow candidates detected now (this time) are considered to be pathologically unrelated.

The positional relationship between the past image and the main image is basically the same as that of the anatomical structure of the subject (thoracic cage, lung field, breast contour, nipple, chest wall, etc.). A correspondence can be obtained by aligning the two images (the past image and the main image) with reference to the position (part) of the structure not to be used.

The part of changing the detection level is that one image to be subjected to abnormal shadow candidate detection processing is
The detection level is not limited to one, but may be two or more, and the detection levels may be different for each part of the plurality. For example, it is preferable to make a difference in the detection level between a portion where the abnormal shadow candidate is easy to detect and a portion where the abnormal shadow candidate is difficult to detect, or to change the detection level for each type of abnormal shadow candidate to be detected.

As an image of a subject in the abnormal shadow candidate detecting method of the present invention, it is practically effective to apply mammography or the like, which is a radiographic image of the breast obtained by imaging for breast cancer screening. It is not necessarily limited to mammography.

As a specific method for detecting an abnormal shadow candidate, it is preferable to apply, for example, a detection method using the above-described abnormal shadow candidate detection system (Japanese Patent Laid-Open No. 8-294479). This abnormal shadow candidate detection method automatically selects a tumor shadow candidate (an aspect of an abnormal shadow candidate) in an image by an iris filter process that evaluates the degree of concentration of a density (signal value) gradient vector of a digital image signal representing an image. Or a morphological operation (a dilation process, an erosion process, an opening process, a closing process, etc.) performed on a digital image signal to automatically detect a microcalcification shadow candidate (an aspect of an abnormal shadow candidate). The detection level can be changed by changing a threshold (e.g., a threshold T1 for comparing and comparing an output value I of an iris filter process described later) in evaluating the degree of concentration of the density gradient vector, In addition, the size of the structural element in the morphological operation is changed or It is possible to change the detection level by changing the threshold value used in degrees evaluation (such as threshold T2, T3 in the later-described equation (17)). Of course, various detection methods other than the abnormal shadow candidate detection method may be applied.

The iris filter processing is performed by Obata et al. Of Tokyo University of Agriculture and Technology, "Detection of tumor shadow in DR image (iris filter)" (Transactions of the Institute of Electronics, Information and Communication Engineers, D-II Vol. J75-D-).
II No.3 P663-670 March 1992), which has been studied as an effective method to detect candidates for tumor shadows, which is one of the characteristic features of breast cancer. The target image is not limited to such a tumor shadow in mammography, but may be applied to any image portion where the gradient of the image signal (density etc.) representing the image is concentrated. This is a possible detection process.

The outline of the process of detecting a tumor shadow candidate using the iris filter will be described below.

For example, in a radiographic image (an image represented by an image signal of high density and high signal level) on an X-ray film, it is generally known that the density of a tumor shadow is slightly lower than that of a surrounding image portion. The density value distribution has a density value gradient such that the density value decreases from the substantially circular periphery toward the center. Therefore, in the tumor shadow, a local gradient of the density value is recognized, and the gradient line is concentrated toward the center of the tumor.

The iris filter calculates the gradient of the image signal represented by the density value as a gradient vector and outputs the degree of concentration of the gradient vector. Based on this, a candidate for a tumor shadow is detected.

That is, for example, in the mammography P as shown in FIG. 7A, the gradient vector at an arbitrary pixel in the tumor shadow PJ is directed to the vicinity of the center of the tumor shadow as shown in FIG. In the case of an elongated shadow PK such as a mammary gland, the gradient vector does not concentrate on a specific point as shown in FIG. If the extracted region is extracted, it becomes a candidate considered as a tumor shadow. It should be noted that, as shown in FIG. 4D, with regard to a shadow PL in which elongated shadows such as a mammary gland intersect, the gradient vector tends to concentrate on a specific point, and may be erroneously detected as a candidate for an abnormal shadow. The above is the basic concept of the iris filter processing. The specific algorithm steps are shown below.

(Step 1) Calculation of Gradient Vector For all the pixels constituting the target image, for each pixel j, the direction θ of the gradient vector of the image data is calculated based on the following equation (1). Ask.

[0027]

(Equation 1) Here, as shown in FIG. 8, f _{1 to} f ₁₆ are pixel values (image data) corresponding to the pixels on the outer periphery of the mask of 5 × 5 pixels centering on the pixel j.

(Step 2) Calculation of Concentration of Gradient Vector Next, for all the pixels constituting the target image,
For each pixel, the degree of concentration C of the gradient vector having that pixel as the pixel of interest is calculated according to the following equation (2).

[0029]

(Equation 2) Here, N is the number of pixels existing in a circle having a radius R around the pixel of interest, θj is a straight line connecting the pixel of interest and each pixel j in the circle, and the above equation (1) for each pixel j (See FIG. 9). Therefore, the degree of concentration C represented by the above equation (2) becomes a large value when the direction of the gradient vector of each pixel j is concentrated on the target pixel.

By the way, the gradient vector of each pixel j in the vicinity of the tumor shadow is determined regardless of the contrast of the tumor shadow.
Since the pixel of interest is directed substantially toward the center of the tumor shadow, the pixel of interest having a large value of the degree of concentration C can be said to be the pixel at the center of the tumor shadow. On the other hand, in the shadow of a linear pattern such as a blood vessel, the value of the degree of concentration C is small because the direction of the gradient vector is biased in a certain direction. Therefore, the value of the degree of concentration C with respect to the pixel of interest is calculated for each of all the pixels constituting the image, and whether or not the value of the degree of concentration C exceeds a predetermined threshold value is evaluated. Can be detected. In other words, this filter is less susceptible to blood vessels and mammary glands than a normal differential filter,
It has the feature that tumor shadows can be detected efficiently.

Further, in the actual processing, in order to achieve a detection power independent of the size and shape of the tumor, a device for adaptively changing the size and shape of the filter is used. FIG. 10 shows the filter. This filter differs from the filter shown in FIG.
M kinds of directions for each π / M degree (11.25 in FIG. 10)
The evaluation of the degree of concentration is performed only with pixels on a radial line (for example, 32 directions for each degree).

Here, the coordinates ([x], [y]) of the n-th pixel on the i-th line and from the pixel of interest are as follows if the coordinates of the pixel of interest are (k, l): Equation (3),
Given by (4).

[0033]

(Equation 3) Here, [x] and [y] are the maximum integers not exceeding x and y.

Further, the output value up to the pixel at which the maximum concentration is obtained for each line on the radial line is defined as the concentration Cimax in that direction, and the concentration Cimax is averaged in all directions to obtain the target value. The degree of concentration C of the gradient vector group for the pixel is set.

Specifically, first, the concentration Ci (n) obtained from the pixel of interest to the n-th pixel on the i-th radial line is obtained by the following equation (5).

[0036]

(Equation 4) That is, equation (5) is based on the target pixel as the starting point and the end point as R
The concentration Ci (n) is calculated within the range from min to Rmax. Here, Rmin and Rmax are the minimum and maximum values of the radius of the tumor shadow to be extracted.

Next, the degree of concentration C of the gradient vector group is calculated by the following equations (6) and (7).

[0038]

(Equation 5) Here, since Cimax in Expression (6) is the maximum value of the concentration Ci (n) for each radial line obtained in Expression (5), the concentration Ci (n) is the maximum value from the pixel of interest. The region up to the pixel becomes a candidate region of the tumor shadow in the direction of the line.

Equation (6) is calculated for all radial lines to determine the area of the tumor shadow on each line, and the area of the tumor shadow on each line is connected by a straight line or non-linear curve between adjacent lines. The shape of the outer peripheral edge of a region that can be a candidate for a tumor shadow can be specified.

In the equation (7), the maximum value Cimax of the degree of concentration given by the equation (6) in this region is calculated for all directions of the radial line (the equation (7) exemplifies the case of 32 directions). Find the average value. The obtained value is the output value I of the iris filter processing, and this output value I is compared with a predetermined fixed threshold value T1 suitable for determining whether or not the shadow is a tumor, and I ≧ T1 If (or I> T1), the area around the target pixel is an abnormal shadow candidate (tumor shadow candidate), and I <T1 (or I ≦ T1).
If 1), it is determined that it is not a tumor shadow candidate.

The manner in which the size and shape of the area for evaluating the degree of concentration C of the gradient vector group in equation (7) adaptively changes in accordance with the distribution of the gradient vector is enlarged in accordance with the brightness of the outside world. Since the iris of the human eye that shrinks is similar to the appearance, the above-described method of detecting a candidate region of a tumor shadow using the degree of concentration of the gradient vector is called iris filter processing. I have.

For the calculation of the concentration Ci (n), the following equation (5 ') may be used instead of the equation (5).

[0043]

(Equation 6) That is, equation (5 ') calculates the degree of concentration Ci (n) from the pixel corresponding to the minimum value Rmin of the radius of the tumor shadow to be extracted to the end point within the range from Rmin to Rmax. .

According to the above-described steps, the iris filter can effectively detect only a tumor shadow of a desired size from a radiographic image.

Generally, the shadow of a malignant mass has the following morphological features: 1) its edge is irregular; 2) it has a shape close to a circle; and 3) the inside has an uneven density distribution. I have.

Therefore, for a more definitive diagnosis, the image signal of the abnormal shadow candidate obtained by the comparison processing between the iris filter output value I and the threshold value T1 described above is subjected to a shape considering these characteristics. The determination may be further performed. The feature amounts used here are the spread degree (Spreadness), the elongation (Elongation), the roughness of the edge (Roughness), the circularity (Circularity), and the inner roughness (Entropy). The final determination as to whether or not the candidate is a tumor shadow may be made by comparing with another predetermined threshold T2 set in advance.
The additional detection processing such as the detection processing based on the morphological features is not the iris filter processing itself, but is applied to the abnormal shadow candidate detection processing by the iris filter processing to detect the abnormal shadow candidate. Since this is a process, these additional detection processes are also included in the iris filter process and are referred to as “process based on the iris filter process”.

On the other hand, the morphology calculation process is a method for detecting a candidate for a microcalcification image, which is a characteristic form of breast cancer, together with a tumor shadow, and uses a multi-scale λ and a structural element (mask) B, and (1) It is effective in extracting the calcified image itself, (2) is not easily affected by complicated background information, and (3) the extracted calcified image is not distorted. In other words, this method is different from general differential processing in that the size, shape,
It is possible to perform detection while better maintaining geometric information such as a density distribution. The outline is described below.

(Basic calculation of morphology) The morphology calculation process is generally developed as a set theory in an N-dimensional space. However, for intuitive understanding, a description will be given of a two-dimensional gray image.

The point of coordinates (x, y) is represented by a density value f
It is regarded as a space having a height corresponding to (x, y). Here, the density value f (x, y) is a high-luminance high-signal level signal which becomes a larger image signal as the density is lower (the brightness is higher when displayed on a CRT).

First, for simplicity, consider a one-dimensional function f (x) corresponding to the cross section. The structuring element g used in the morphological operation processing is a symmetric function symmetric about the origin as shown in the following equation (8).

(Equation 7) And the value is 0 in the domain, and the domain is represented by the following equation (9).

[0051]

(Equation 8) At this time, the basic form of the morphological operation is expressed by equations (10) to (1).
As shown in 3), this is a very simple operation.

[0052]

(Equation 9) That is, the dilation process is a process of searching for a maximum value within a range of ± m (a value determined according to the structural element B) around the target pixel (FIG. 11).
On the other hand, the erosion process is a process of searching for a minimum value within a width of ± m around the pixel of interest (see FIG. 11B). The opening processing corresponds to searching for the maximum value after searching for the minimum value, and the closing processing corresponds to searching for the minimum value after searching for the maximum value. The opening process is performed from the low luminance side to the density curve f (x).
Is equivalent to removing a convex density fluctuation portion (a portion higher in luminance than the surrounding portion) which fluctuates in a spatially narrower range than the mask size 2 m (see FIG. 11C). On the other hand, in the closing processing, the density curve f (x) is smoothed from the high brightness side, and a concave density variation portion (a portion having a lower brightness than the surrounding portion) which fluctuates in a space narrower than the mask size 2 m is removed. (FIG. 11
(D)).

Here, in the case of a high-density high-signal level signal in which the higher the density, the higher the density, the density value f
Since the magnitude relationship is reversed with respect to the case where the image signal value of (x) is at the high luminance and high signal level, the dilation processing at the high density and high signal level is performed by the erosion processing at the high luminance and high signal level (FIG. B)), the erosion processing on the signal of high density and high signal level
Dilation processing at high luminance and high signal level (Fig. 1
1 (A)), and the opening process for a signal with a high density and high signal level matches the closing process for a signal with a high luminance and high signal level (FIG. 11D). Is an opening process at a high luminance and a high signal level (FIG. 11C).
Matches. In this section, a case of an image signal (luminance value) at a high luminance and high signal level will be described.

(Application to Calculated Shading Detection) For the detection of calcified shading, a difference method of subtracting a smoothed image from an original image can be considered. It is difficult to distinguish between calcified shadows and elongated non-calcified shadows (e.g., mammary glands, blood vessels, and mammary gland supporting tissues) using a simple smoothing method. Obata et al. We propose a morphological operation represented by equation (14) ("Extraction of microcalcification image by morphological filter using multiple structural elements" IEICE Transactions D-II Vol.J75-D-II No.
7 P1170-1176, July 1992, "Basics of Morphology and Its Application to Mammogram Processing" MEDICALIMAGING TEC
HNOLOGY Vol.12 No.1 January 1994).

[0055]

(Equation 10) Here, Bi (i = 1, 2, 3, 4) is the four linear structural elements B shown in FIG. If the structural element B is set to be larger than the calcified shadow to be detected, the calcification is a convex signal change portion (an image portion that fluctuates in a narrow spatial range) finer than the structural element B in the opening process. The image is removed. On the other hand, the elongated non-calcified shadow has a longer length than the structural element B, and if its inclination (extending direction) matches any of the four structural elements Bi, the opening process (the equation (14)) Even if the two-term operation is performed, it remains as it is. Therefore, by subtracting the smoothed image (image from which the calcified shadow has been removed) obtained by the opening process from the original image f, an image including only small calcified image candidates is obtained. This is the idea of equation (14).

As described above, in the case of a signal having a high density and a high signal level, the calcified shadow has a lower density value than the surrounding image portion, and the calcified shadow has a concave signal change with respect to the surrounding portion. Therefore, the closing process is applied instead of the opening process, and the expression (15) is applied instead of the expression (14).

[0057]

[Equation 11] However, even in this case, some non-calcified shadows having the same size as the calcified shadow may remain. In such a case, the differential information based on the morphological operation of the following equation (16) is used. The non-calcified image included in P in equation (14) is further removed.

[0058]

(Equation 12) Here, as the value of Mgrad is larger, the possibility of calcification shadow is larger, and the calcification candidate image Cs can be obtained by the following equation (17).

[0059]

(Equation 13) Here, T2 and T3 are preset thresholds determined experimentally.

However, for non-calcified shadows different from the size of the calcified shadow, P in equation (14) and a predetermined threshold T2
In the case where a non-calcified shadow having the same size as the calcified shadow does not remain since it can be removed only by comparing with the condition (P (i, j) .Gtoreq.T2).

As described above, these morphological operation processes are described for the case of image data of high luminance and high signal level. However, the image data of high density and high signal level (the higher the density of pixels, the larger the digital value is). ), The opening operation and the closing operation have the opposite relationship.

In the morphology calculation processing, additional detection processing similar to additional detection processing such as detection processing based on morphological features in the iris filter processing may be further applied. Such additional detection processing is also included in the morphology calculation processing, and is referred to as “processing based on morphology calculation processing”.

When the above-described abnormal shadow candidate detection processing is applied, the threshold T1 to be compared with the iris filter output value I, the morphology operation output value P (i, j), Changing the threshold values T2 and T3, which are respectively compared with Mgrad, is as follows:
Although simple and preferable, the present invention is not limited to this mode. For example, the size of a structural element used for morphological operation may be changed, or another parameter may be changed. When the candidate detection processing is applied, in the detection processing, a threshold value, a parameter, or the like that affects the detection rate of the abnormal shadow candidate may be changed.

An abnormal shadow candidate detecting system according to the present invention is an apparatus for implementing the abnormal shadow candidate detecting method according to the present invention, and detects an abnormal shadow candidate in an image of a subject based on the image of the subject. In an abnormal shadow candidate detection system provided with candidate detecting means, a past information obtaining means for obtaining information on detection of a past abnormal shadow candidate for a specific portion of the subject, and the past obtained by the past information obtaining means. In accordance with the information on the detection of abnormal shadow candidates, the image further comprises a detection level changing means for changing a detection level of at least a part corresponding to the specific portion, the abnormal shadow candidate detecting means,
The abnormal shadow candidate is detected based on the detection level changed by the detection level changing unit.

The past information acquiring means includes, for example, an I-number assigned to each subject or each patient who is the subject.
A past abnormality corresponding to an ID number of a patient to be detected this time, which is input from a storage medium in which information on the detection of a past abnormal shadow candidate is stored in association with a D number or the like. What reads information about the detection of a shadow candidate may be applied.

The detection level changing means stores the detection level in advance in a look-up table system in accordance with the content of the information on the detection of the past abnormal shadow candidate, and corresponds to the acquired information on the detection of the past abnormal shadow candidate. What is necessary is to apply the detection level to be input to the abnormal shadow candidate detection means.

For example, as the detection level changing means, when the information relating to the detection of the abnormal shadow candidate indicates that the abnormal shadow candidate has been detected in the past for the specific portion, the abnormal shadow candidate Preferably, the detection level for at least a portion is changed so as to increase the detection rate.

In the abnormal shadow candidate detecting system according to the present invention, the detection level changing means for changing the detection level for each part of the image is applied, and the abnormal shadow candidate detecting means is changed for each part by the detection level changing means. It is preferable to apply a method for detecting an abnormal shadow candidate based on the detection level changed to.

[0069]

According to the abnormal shadow candidate detecting method and the abnormal shadow candidate detecting system of the present invention, the detection level of the abnormal shadow candidate is changed for at least a part according to the information on the detection of the past abnormal shadow candidate for the subject. By doing so, it is possible to easily detect an abnormal shadow candidate in the part, thereby suppressing omission in detection, or making it difficult to detect, thereby suppressing erroneous detection, and consequently increasing the reliability of detection as compared with the conventional method. Can be.

[0070]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, specific embodiments of the abnormal shadow candidate detecting method and abnormal shadow candidate detecting system of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of a first abnormal shadow candidate detection system of the present invention for implementing the first abnormal shadow candidate detection method of the present invention, and FIG. FIG. 1 is a block diagram showing a computer-aided image diagnostic apparatus 20 incorporating a shadow candidate detection system 10. The illustrated computer-aided image diagnostic apparatus 60 receives an input image signal (hereinafter, referred to as an image signal).
A storage means 20 for storing an entire image signal (hereinafter referred to as an entire image signal) S, and an entire image obtained by reading out the entire image signal S stored in the storage means 20 and subjecting the entire signal S to image processing such as gradation processing and frequency processing The processing unit 30 and the whole image signal S stored in the storage unit 20 are read, and image signals (hereinafter, referred to as local image signals) Sp and Sq representing the abnormal shadow candidates Pp and Pq are detected from the whole image signal S. An abnormal shadow candidate detection system 10, a local image processing unit 40 for emphasizing an abnormal shadow candidate Pp, Pq with respect to the detected local image signals Sp, Sq, and an entire image processing unit 30
, And an abnormal shadow candidate Pp ′, represented by local image (abnormal shadow candidate) signals Sp ′, Sq ′ represented by the whole image signal S ′ after image processing by the local image processing means 40. And an image display means 50 for displaying Pq ′ as a visible image. Note that the code Pp
Represents a tumor shadow candidate, and a symbol Pq represents a microcalcification shadow candidate, respectively, and symbols Sp and Sq are signals corresponding to these, respectively.

Here, the whole image signal S input to the computer-aided image diagnostic apparatus 60 is, for example, a stimulable fluorescent light in which a mammography P representing a breast radiation image of a patient as shown in FIG. This is an image signal (a signal with a high density and a high signal level) obtained by photoelectrically reading the photostimulated light emitted by irradiating the body sheet with the excitation light, and then performing digital conversion.

The image display means 50 may display the entire image P and the abnormal shadow candidates Pp 'and Pq' on the display surface separately, but in the present embodiment, the entire image P is displayed. Only the image portions of the abnormal shadow candidates Pp and Pq in the whole image P are replaced with the abnormal shadow candidate images Pp ′ and Pq ′ processed by the local image processing means 40 and displayed.

As shown in detail in FIG. 1, the abnormal shadow candidate detection system 10 includes a past information obtaining means 2 for obtaining information on abnormal shadow candidates detected in the past in a specific portion of a subject, and a past information obtaining means 2. According to the information about the abnormal shadow candidate detected in the past obtained by
Detection level changing means 3 for changing the detection level of the portion of the image P corresponding to the specific portion so as to increase the detection rate of the abnormal shadow candidate, and detecting the abnormal shadow candidates Pp and Pq. Is applied to the entire image signal S, and the output value of the obtained arithmetic processing and the preset threshold value T (collectively referred to as T1, T2, and T3 described below) ), And for the above part, the output value of the arithmetic processing obtained is compared with a threshold value T ′ corresponding to the detection level changed by the detection level changing means 3 so as to obtain a comparison result. An abnormal shadow candidate detecting means 1 for detecting and processing abnormal shadow candidates Pp and Pq.

The object to be diagnosed by the computer-aided image diagnostic apparatus 60 shown in FIG. 2 is a mammography (whole image) P obtained by a breast cancer screening, and the abnormal shadow is a tumor which is a mode of breast cancer. Shading Pp and microcalcification shading P
q (see FIG. 3A).

The detection of the abnormal shadow candidates Pp and Pq by the abnormal shadow candidate detecting means 1 is performed by the iris filter processing described above and the subsequent iris filter output value I and the threshold value T1.
And a microcalcification shadow candidate detection consisting of the morphology calculation processing described above and the threshold processing with thresholds T2 and T3 in equation (17) described above. If the iris filter output value I of the predetermined area is larger than the threshold value T1 (I> T1), the area is detected as a tumor shadow candidate Pp, and if the iris filter output value I is equal to or smaller than the threshold value T1 (I ≦ T1), the area is designated as a tumor shadow candidate Pp
, And the result of the morphological operation P (i,
j) and Mgrad (see equations (15) and (16)) are equal to or greater than thresholds T2 and T3, respectively (P (i, j) ≧ T2 and Mgrad ≧ T3), and the corresponding area is designated as a microcalcification shadow candidate Pq. If the result P (i, j) or Mgrad of the morphological operation is smaller than the threshold values T2 and T3, (P (i,
j) <T2 or Mgrad <T3), a process of not detecting the region as a microcalcification shadow candidate Pq. Therefore, the detection level of these tumor shadow candidates is equal to the threshold T1.
, And the detection level of the microcalcification shadow candidate can be changed by changing the thresholds T2 and T3.

The information relating to the abnormal shadow candidate detected in the past, which is input to the past information acquiring means 2, is, for example, as shown in FIG.
As shown in (2), this represents the detection result of the abnormal shadow candidate based on the past image P of the patient, and the regions P1 and P3 in which the tumor shadow candidate Pp is detected, and the microcalcification shadow candidate Pq This is information indicating the detected area P2. Therefore, the detection level changing means 3 corresponds to the areas P1 and P3 in which the tumor shadow candidates Pp acquired by the past information acquiring means 2 are detected, respectively, as shown in FIG.
Regarding the region in the image of the current detection processing target shown in (1), the threshold T1 is reduced, and the region in the image of the current detection processing target corresponding to the region P2 in which the microcalcification shadow candidate Pq is detected is described. Is to change the detection level so as to reduce the threshold values T2 and T3.

Next, the operation of the computer-aided image diagnostic apparatus 60 including the abnormal shadow candidate detecting system 10 of the present embodiment will be described.

The computer-aided image diagnostic apparatus 60
The whole image signal S representing the mammography P shown in (1) is input, and the whole image signal S is temporarily stored in the storage unit 20. Here, the whole image processing means 30 reads out the whole image signal S stored in the storage means 20, and performs image processing such as gradation processing and frequency processing on the whole image signal S. On the other hand, the abnormal shadow candidate detection system 10 also reads out the entire image signal S stored in the storage unit 20, and the entire image signal S is input to the abnormal shadow candidate detection unit 1.

The past information acquiring means 2 of the abnormal shadow candidate detection system 10 includes information on the above-mentioned detection of abnormal past shadow candidates (detection region information and the type of detected abnormal shadow candidate (tumor shadow candidate or minute Whether the candidate is a calcified shadow candidate) is input. Note that these pieces of past detection information are input from a predetermined storage medium or the like stored in association with, for example, an ID number or the like assigned to each patient.

The past detection information input to the past information obtaining means 2 is input to the detection level changing means 3, and the detection level changing means 3 responds to the contents of the input past detection information as described above. In the region corresponding to the regions P1 and P3 with the history where the tumor shadow candidate Pp is detected (corresponding region in the image P shown in FIG. 3A), an abnormal shadow candidate of the type detected in the past (a tumor shadow candidate) The threshold value T1 stored in the abnormal shadow candidate detecting means 1 is changed to be small so that the abnormal shadow candidate detecting means 1 is easily detected, and the threshold value T1 corresponds to the region P2 in the history where the microcalcification shadow candidate Pq has been detected. For the region (corresponding region in the image P shown in FIG. 3A), the type of abnormal shadow candidate (microcalcification shadow candidate) detected in the past is easily detected.
The threshold values T2 and T stored in the abnormal shadow candidate detecting means 1
Change 3 to be smaller.

The abnormal shadow candidate detecting means 1 comprises a storage means 20
The above-described iris filter processing and morphology calculation processing are separately performed on the entire image signal S read from the multiplexed image signal S, and the output value I and the output value I, which represent the signal Sp of the tumor shadow candidate and the signal Sq of the microcalcification shadow candidate, respectively, P (i,
j), Mgrad is calculated. Then, these calculated output values I, P (i, j), and Mgrad are compared with corresponding thresholds T1, T2, and T3, respectively. If the iris filter output value I is larger than the threshold T1, (I> T1) ,
The region is detected as a signal Sp representing the tumor shadow candidate Pp, and if the iris filter output value I is equal to or smaller than the threshold value T1 (I ≦ T1), the region is detected as the tumor shadow candidate Pp.
Is not detected as the signal Sp representing the output signal, and the output values P (i, j) and Mgrad of the morphological operation (Equations (15) and (16))
) Are equal to or greater than thresholds T2 and T3, respectively (P
(I, j) ≧ T2 and Mgrad ≧ T3), the region is detected as a signal Sq representing the microcalcification shadow candidate Pq, and the output value P (i, j) or Mgrad of the morphological operation is smaller than the threshold values T2 and T3. If it is smaller (P (i, j) <T2 or Mgrad <T3), the area is set as a microcalcification shadow candidate Pq
Is not detected as the signal Sq representing At this time, the thresholds T1, T2, T3 are detected by the detection level changing means 3.
Are compared with the changed threshold values T1 ', T2', T3 ', respectively.

As a result, for example, in the mammography P shown in FIG. 3A, according to the original threshold value T1, even if the tumor shadow candidate is not finally detected as a tumor shadow candidate (Pp indicated by a broken line), In the area P1 (FIG. 2B) in which a tumor shadow candidate has been detected in the past, it is detected according to the changed threshold value T1 '. It should be noted that even in the region where the tumor shadow candidate has been detected in the past, the threshold T
Those with an output value I that is lower than 1 ′ (in a region (not shown) corresponding to the region P3 in which the tumor shadow candidate Pp is detected in FIG. 2B) are not detected. The same applies to microcalcification shadow candidates.

The local image signals Sp and Sq respectively representing the tumor shadow candidate Pp (areas P1 and P4) and the microcalcification shadow candidate Pq (area P2) detected by the abnormal shadow candidate detection means 1 are converted to local image processing means, respectively. 40, the local image processing means 40 inputs the local image signal S
Image processing is performed so that abnormal shadow candidates Pp and Pq detected for p and Sq are emphasized, and processed local image signals Sp ′ and Sq ′ are output.

The whole image signal S ′ obtained by processing by the whole image processing means 30 and the local image signals Sp ′ and Sq ′ obtained by processing by the local image processing means 40 are input to the image display means 50. The image display means 50 displays the whole image P ', and only the image part of the abnormal shadow candidate in the whole image P' is an image Pp ', Pq' of the abnormal shadow candidate which has been image-processed by the local image processing means 40. Replaced with and displayed.

As a result, the image display means 50 displays the whole image P 'and the images Pp' and Pq 'of the abnormal shadow candidate emphasized from the whole image P', and diagnoses by an interpreter such as a doctor. (FIG. 4). The displayed abnormal shadow candidate images Pp ′ and Pq ′ are detected after changing the detection level only in the specific region in accordance with the contents of the past detection information on the subject. (In the present embodiment, for example, the region P
In 1), the abnormal shadow candidate can be easily detected, and the abnormal shadow candidate is detected in the past but not detected this time, so that it is possible to prevent the abnormal shadow candidate from being overlooked and forgetting to follow up. As a result, the reliability of detection can be improved as compared with the conventional case.

The abnormal shadow candidate detection system 10 according to the present embodiment changes the threshold value to be compared with the output value of the iris filter process and the output value of the morphology operation process for a specific region, thereby changing the threshold value in the region. Although the detection level of the abnormal shadow candidate is changed, the method of changing the detection level is not limited to the method described in the present embodiment. For example, changing the size or shape of the structural element in the morphological operation processing May change the detection level.

Further, the detection level may be increased by increasing the output value of the iris filter process or the output value of the morphology operation process only for the specific region.

FIG. 5 is a block diagram showing an embodiment of the second abnormal shadow candidate detecting system according to the present invention which implements the second abnormal shadow candidate detecting method according to the present invention. In the abnormal shadow candidate detection system 10 of the embodiment shown in FIG. 1, the detection level changing unit 3 previously changes the detection level of the area in the current image corresponding to the area where the abnormal shadow candidate was detected in the past. However, the abnormal shadow candidate detection system 10 ′ of the embodiment shown in FIG. 5 operates as an abnormal shadow candidate detection unit based on the entire image signal S representing the mammography P and according to the detection level. Is not directly detected, but an expected candidate detecting unit 1a first detects an abnormal shadow prospect candidate that is likely to be detected as an abnormal shadow candidate, and an abnormality detected by the prospect candidate detecting unit based on a predetermined detection level. An abnormal shadow candidate determining unit 1b that determines an abnormal shadow prospect candidate to be finally detected as an abnormal shadow candidate among the shadow prospect candidates; Applying the candidate detecting means 1 ′, as the detection level changing means, information on the abnormal shadow candidates detected in the past (detection area information and the type of the detected abnormal shadow candidates) acquired by the past information acquiring means 2; Then, the region and the type of the abnormal shadow prospect candidate detected by the prospect candidate detecting unit 1a are compared, and in both cases, the abnormal shadow candidate of the same type (in the current target image, the abnormal shadow prospect candidate) exists in substantially the same region. In this case, the detection level is changed so that the threshold T corresponding to the type is reduced for the area, and the original thresholds T1 and T2 stored in the abnormal shadow candidate determination unit 1b are changed for the area other than the area. , T3, to which a detection level changing means is applied.

The prospective candidate detecting unit 1a uses the abnormal shadow candidate detecting means 1 shown in FIG. 1 to set thresholds T1 ", T2", T2 ", T2", which are significantly smaller than the thresholds T1, T2, T3 for detecting abnormal shadow candidates. By T3 ″, potential candidates that can be abnormal shadow candidates are detected. For example, as shown in FIG. 6A, more potential candidates Pp and Pq than those shown in FIG. 3A are detected. You.

Then, the detection level changing means 3 '
Compared with the information on the abnormal shadow candidates detected in the past shown in (2), the areas where the abnormal shadow candidates of the same type (candidate abnormal shadow candidates in the current target image) exist in substantially the same area in both areas (Area P1 of a tumor shadow candidate Pp in the past information, a region P4 of a tumor shadow prospect candidate Pp in the current image, a region P2 of a microcalcification shadow candidate Pq in the past information, and a microcalcification shadow candidate in the current image For the region P7) of Pq, the detection level is changed so as to reduce the threshold value T corresponding to the type. That is, for the region P4, the threshold value T that is smaller than the original threshold value T1 and that is used when the prospective candidate Pp is detected.
1 ″ (T1> T1 ′ ≧ T
1 ″), and for the other areas, the original threshold T1.
Further, for the region P7, the thresholds T smaller than the original thresholds T2 and T3 and equal to or larger than the thresholds T2 ″ and T3 ″ when the prospective candidate Pq is detected, respectively.
2 ′, T3 ′ (T2> T2 ′ ≧ T2 ″, T3> T3 ′ ≧
T3 ″) and the original threshold values T2 and T3 for the other areas.
The detection level of the abnormal shadow candidate by the abnormal shadow candidate determining unit 1b is changed.

The abnormal shadow candidate determining section 1b determines whether a specific area P
4 and P7, the detection levels changed by the detection level changing means 3 '(the area P4 has thresholds T1', T2, T
3. In the area P7, according to the thresholds T1, T2 ', T3'), and in the other areas, the original detection level (threshold T)
(1, T2, T3), threshold processing is performed on each of the areas P4 to P10 detected by the prospective candidate detection unit 1a. As a result, for example, the abnormal shadow candidates Pp and Pq in the regions P4, P7, and P8 respectively exceed the threshold values of the corresponding types, and are determined as final abnormal shadow candidates.

Here, according to the original threshold value T1, for example, the abnormal shadow prospect candidate Pp in the area P4 is not finally detected as a tumor shadow candidate, but it is a tumor shadow candidate of the same type in the past. Is detected in the area corresponding to the detected area P1, and thus is detected by the changed threshold value T1 '.

The local image signals Sp and Sq respectively representing the tumor shadow candidate Pp (areas P4 and P8) and the microcalcification shadow candidate Pq (area P7) detected by the abnormal shadow candidate detecting means 1 'are respectively processed by local image processing. Input to the local image processing means 40, and the local image processing means 40 detects abnormal shadow candidates Pp, Pq detected for the input local image signals Sp, Sq.
Are processed so as to be emphasized, and the processed local image signals Sp ′ and Sq ′ are output.

The whole image signal S ′ obtained by processing by the whole image processing means 30 and the local image signals Sp ′ and Sq ′ obtained by processing by the local image processing means 40 are inputted to the image display means 50. The image display means 50 displays the whole image P ', and only the image part of the abnormal shadow candidate in the whole image P' is an image Pp ', Pq' of the abnormal shadow candidate which has been image-processed by the local image processing means 40. Replaced with and displayed.

As a result, the image display means 50 displays the image Pp 'and Pq' of the abnormal shadow candidate emphasized from the whole image P 'together with the whole image P', and diagnoses by a doctor or the like. To be served. Then, the displayed abnormal shadow candidate images Pp ′ and Pq ′ have been detected by changing the detection level only in the specific region in accordance with the contents of the past detection information on the subject, so that the region ( In the present embodiment, for example, an abnormal shadow candidate can be easily detected in the region P4). Since the abnormal shadow candidate was detected in the past but was not detected this time, the abnormal shadow candidate was overlooked, and subsequent follow-up observation was performed. Forgetting or forgetting can be prevented, and as a result, the reliability of detection can be improved as compared with the related art.

The abnormal shadow candidate detection system 10 'of the present embodiment uses the computer-aided image diagnostic apparatus 60 shown in FIG.
Can be replaced with the abnormal shadow candidate detection system 10.

In the abnormal shadow candidate detection system of each of the above-described embodiments, only the detected part and the type of the abnormal shadow candidate are applied as information on the abnormal shadow candidate detected in the past. In the candidate detection method and the detection system, not only those of this aspect, but also include the degree of malignancy and size of the detected abnormal shadow candidate, the necessity of follow-up, the diagnosis information of the doctor, etc. The detection level changing means may change the detection level more finely according to the additional past information.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram showing an embodiment of a first abnormal shadow candidate detection system of the present invention.

FIG. 2 is a block diagram showing an example of a computer-aided image diagnostic apparatus using the abnormal shadow candidate detection system shown in FIG.

FIG. 3 is a diagram showing a mammography P represented by an image signal S input to the computer-aided image diagnostic apparatus shown in FIG. 3;

FIG. 4 is a diagram showing an image displayed and output by the computer-aided image diagnostic apparatus shown in FIG. 1;

FIG. 5 is a schematic block diagram showing an embodiment of a second abnormal shadow candidate detection system of the present invention.

FIG. 6 is a view showing a mammography P represented by an image signal S input to the computer-aided image diagnostic apparatus shown in FIG. 5;

FIG. 7 is a conceptual diagram showing the degree of concentration of a density gradient in mammography.

FIG. 8 is a diagram showing a mask for calculating a gradient vector in the iris filter processing.

FIG. 9 is a diagram showing the concept of the degree of concentration of a gradient vector for a target pixel.

FIG. 10 is a conceptual diagram showing an iris filter set so that a contour shape is adaptively changed.

FIG. 11 is a graph illustrating a basic operation of the morphological operation processing;

FIG. 12 is a diagram showing the concept of a structural element used in morphological operation processing.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Abnormal shadow candidate detection means 2 Past information acquisition means 3, 3 'Detection level changing means 10 Abnormal shadow candidate detection system 20 Storage means 30 Overall image processing means 40 Local image processing means 50 Display means 60 Computer-aided image diagnostic apparatus

 ────────────────────────────────────────────────── ─── Continuing from the front page (72) Inventor Takeshi Okubo Takeshi, 798-ga, Kaisei-cho, Ashigaruga-gun, Kanagawa Fuji Photo F Co., Ltd. F-term (reference) 4C093 AA26 CA18 CA35 DA06 FF08 FF09 FF17 FF19 FF28 5B057 AA08 BA03 BA24 BA30 CC03 CE05 CE11 CH11 DA07 DA08 DC32 5C054 AA01 AA07 CA02 CC01 EA05 ED11 FA01 FA02 FB01 FC03 FC12 FE09 FE12 HA12

Claims (12)

[Claims] 1. An abnormal shadow candidate detecting method for detecting an abnormal shadow candidate in an image based on an image of a subject, wherein the abnormal shadow candidate detecting method includes the steps of: An abnormal shadow candidate detection method, comprising: changing a detection level of at least a part of an image corresponding to the specific part. 2. When the information relating to the detection of the abnormal shadow candidate indicates that the abnormal shadow candidate has been detected in the past for the specific portion, the detection rate of the abnormal shadow candidate in the portion is increased. The abnormal shadow candidate detecting method according to claim 1, wherein the detection level of at least a part is changed as described above. 3. The abnormal shadow candidate detection according to claim 1, wherein the information on the detection of the past abnormal shadow candidate includes at least information for specifying the position of the abnormal shadow candidate in the subject. Method. 4. The information relating to the detection of the past abnormal shadow candidate is at least one of the size, type, malignancy, and diagnostic information of the abnormal shadow candidate.
The abnormal shadow candidate detecting method according to claim 3, further comprising: 5. The abnormal shadow candidate detection method according to claim 1, wherein the detection level is changed for each part of the image. 6. The abnormal shadow candidate detecting method according to claim 1, wherein the image of the subject is a mammography. 7. An abnormal shadow candidate detecting system comprising an abnormal shadow candidate detecting means for detecting an abnormal shadow candidate in an image based on an image of the object, wherein an abnormal shadow candidate is detected for a specific portion of the object. Past information acquiring means for acquiring information about the detection, and according to the information about the detection of the past abnormal shadow candidate acquired by the past information acquiring means, at least a part of the image corresponding to the specific part. A detection level changing unit that changes a detection level, wherein the abnormal shadow candidate detecting unit detects the abnormal shadow candidate based on the detection level changed by the detection level changing unit. Abnormal shadow candidate detection system. 8. The detection method according to claim 1, wherein the detection level changing unit determines that the abnormal shadow candidate is detected in the part in a case where the information relating to the detection of the abnormal shadow candidate indicates that the abnormal shadow candidate has been detected in the past. The abnormal shadow candidate detection system according to claim 7, wherein a detection level of at least a part of the system is changed so as to increase a candidate detection rate. 9. The abnormal shadow candidate detection according to claim 7, wherein the information relating to the detection of the past abnormal shadow candidate includes at least information for specifying the position of the abnormal shadow candidate in the subject. system. 10. The information relating to the detection of a past abnormal shadow candidate further includes at least one of the size, type, malignancy, and diagnostic information of the abnormal shadow candidate. Claim 9
The described abnormal shadow candidate detection system. 11. The abnormal shadow candidate according to claim 7, wherein the detection level changing unit changes the detection level for each part of the image. Detection system. 12. The abnormal shadow candidate detection system according to claim 7, wherein the image of the subject is a mammography.