JP2012255808A - Pathologic tissue image imaging system, pathologic tissue image imaging method, and pathologic tissue image imaging program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To photograph a pathologic tissue image, by detecting a region (an ROI) in which a pathologist is interested from the pathologic tissue image in histopathology diagnosis without operation by the pathologist, by selecting the detected ROI as a range to be expanded and photographed, and by expanding and photographing the selected range.SOLUTION: A pathologic tissue image imaging system comprises: a histopathology image acquiring section for photographing a pathologic tissue image; an outputting section for outputting the pathologic tissue image; a weighting section for detecting an ROI from the pathologic tissue image for adding weight to a pixel located in the ROI; a range selecting section for selecting an expanded photographing range in which the pathologic tissue image is expanded and photographed on the basis of a weight value that is added by the weighting section. The histopathology image acquiring section photographs the pathologic tissue image that is expanded in the expanded photographing range selected by the range selecting section. The outputting section outputs the pathologic tissue image that is expanded in the expanded photographing range and is photographed.

Description

  The present invention selects a range to be magnified and photographed from a pathological tissue image, and photographs a pathological tissue image in which the selected range is magnified, a pathological tissue image photographing method, and a pathological tissue image photographing program , Regarding.

  In recent clinical practice, diagnostic imaging equipment such as X-ray, CT (Computed Tomography), MRI (Magnetic Resonance Imaging) and the like have advanced, and minute foreign substances in the body can be detected. However, these diagnostic imaging devices only provide position information of foreign bodies in the body and cannot identify the nature of the foreign bodies.

  Therefore, the specimen tissue collected from the foreign object detected by the pathologist with the diagnostic imaging device is finally observed with a microscope, and the diagnosis is based on experience whether the detected foreign substance is benign or malignant (pathological tissue diagnosis )is doing.

  In particular, histopathological diagnosis of cancer is performed as follows.

  First, a specimen tissue collected from a foreign substance is dehydrated to fix it and subjected to a blocking process with paraffin.

  Next, the block-processed specimen tissue is cut into thin slices having a thickness of 4 to 8 micrometers and placed on a slide glass to create a pathological tissue slide.

  Next, paraffin is removed from the specimen tissue on the prepared pathological tissue slide and stained with a staining solution of hematoxylin and eosin (HE staining). By HE staining, cell nuclei contained in the specimen tissue are stained blue-purple, and other cytoplasm, fibers, erythrocytes, etc. are stained light red.

  Thereafter, the pathologist observes the HE-stained pathological tissue slide with a microscope and performs histopathological diagnosis based on the morphological information obtained from the observation results.

  At this time, the pathologist first observes the specimen tissue on the pathological tissue slide through a low-magnification objective lens, and changes the pattern of the specimen tissue as a tissue (for example, changes in the density of tissue and cell nuclei, Observe pattern changes, etc.) and highlight the areas suspected of being cancer based on the observations.

  After that, the pathologist observes the suspected cancer area through a high-magnification objective lens, observes changes in the size and shape of the cell nucleus of the specimen tissue, and examines the nature of the foreign material from which the specimen tissue was collected. judge.

  In recent years, pathological doctors photograph a pathological tissue slide as a pathological tissue image and perform a diagnosis based on the photographed pathological tissue image during histopathological diagnosis.

  By taking the pathological tissue slide in the microscopic field as a pathological tissue image, it can be easily compared with the pathological tissue images taken in the past and typical pathological tissue images of each disease, so the pathologist is more accurate. Pathological diagnosis can be performed.

  Recently, the pathological tissue images taken are analyzed, and based on the morphological information of the specimen tissue obtained from the histopathological image, the diagnostic category to which the specimen tissue belongs and the malignancy of cancer cells contained in the specimen tissue. Development of a pathological tissue diagnosis support system and the like that supports pathological diagnosis by a pathologist by performing simple histopathological diagnosis such as determining the As an example of a histopathological diagnosis support system, Patent Document 1 describes quantitatively the features of a histopathological image and calculates the fitness of each diagnosis category based on the histopathological features. Techniques for displaying high diagnostic category names are described.

  In general, a histopathological image used for such a histopathological diagnosis is taken by a CCD (Charge Coupled Device) camera mounted on a microscope.

  Previously, it was necessary to develop the photographed pathological tissue image as a photograph. Recently, it has become possible to confirm the photographed pathological tissue image on a monitor, saving time for developing the photographed pathological tissue image as a photograph. The time required for diagnosis can be shortened.

  In addition, by saving the captured pathological tissue image as electronic information, comparison with a previously captured pathological tissue image, image processing on the pathological tissue image, etc. can be performed more easily, and rapid pathological tissue Diagnosis can be performed.

  Various types of such pathological tissue imaging systems that capture a pathological tissue image and output the captured pathological tissue image to a monitor have been proposed. There is also a technique for assisting pathological diagnosis by a pathologist by performing processing such as processing.

  As an example, Patent Document 2 discloses that a pathological tissue image photographed at a low magnification is divided into several small sections, the ratio of the tissue area included in each small section is calculated, and the ratio of the tissue area is high. A technique is described in which a priority is given to small sections in order to present a guideline when a pathologist determines a region of interest (ROI: Region Of Interest) in histopathological diagnosis.

JP 2006-153742 A Japanese Patent Laid-Open No. 11-344676

  However, the technique described in Patent Document 2 has the following problems.

  The first problem is that the ratio of the tissue region occupying the photographed pathological tissue image does not indicate the importance of the range as it is.

  For example, when a tissue exists in almost the entire area of a pathological tissue slide, such as a breast cancer mammary gland surgery material, the entire pathological tissue image obtained by photographing the pathological tissue slide is occupied by the tissue.

  In the case of cancer, since its characteristics are likely to appear in the cell nucleus, the range including many cell nuclei is often important. For this reason, when the entire pathological tissue image is occupied by the tissue, the pathologist observes the region including many cell nuclei as an ROI in an enlarged manner.

  On the other hand, even if the ratio of the tissue area that occupies the pathological tissue image is high, if the area is only fat or stroma, the range is less important. We do not pay attention to such areas.

  In the technique described in Patent Document 2, a taken pathological tissue image can be divided into small sections, and priorities can be provided in descending order of the proportion of the tissue region. However, as described above, the range in which the ratio of the tissue area is high does not necessarily match the ROI. Therefore, the priority assigned to each small section by the technique described in Patent Document 2 does not always become a priority when the pathologist enlarges and captures the ROI as it is.

  Therefore, from a pathological tissue image, an ROI that is an area that a pathologist pays attention to when performing histopathological diagnosis, such as an area that includes a large number of cell nuclei, is detected instead of simply detecting an area that is occupied by a tissue area. There is a need.

  The second problem is that the technique described in Patent Document 2 cannot select an enlarged photographing range that is highly important and magnified unless it is operated by a pathologist.

  As described in the first problem, the priority assigned by the technique described in Patent Document 2 does not necessarily coincide with the priority at the time when the pathologist enlarges and takes an image as the ROI. Therefore, in the technique described in Patent Document 2, it is necessary for the pathologist to check the pathological tissue image captured on the monitor while referring to the priority assigned by the pathological tissue imaging system and select the enlarged imaging range. is there.

  In addition, since the enlarged imaging range cannot be selected unless the pathologist operates the technique described in Patent Document 2, the pathological tissue diagnosis support system and the pathological tissue imaging system described in Patent Document 1 are provided. It cannot be incorporated into an automated pathological tissue diagnosis support system that automatically performs all processes from taking a pathological tissue image to pathological tissue diagnosis.

  Therefore, it is necessary to automatically detect the ROI from the pathological tissue image, select the detected ROI as the enlarged imaging range, and capture the pathological tissue image in which the selected enlarged imaging range is enlarged.

  In summary, in order to solve the above first and second problems, the ROI is automatically detected from the pathological tissue image, the detected ROI is selected as the enlarged imaging range, and the selected enlarged imaging range is selected. There is a problem that it is necessary to take an enlarged pathological tissue image.

  An object of the present invention is to provide a pathological tissue image capturing system, a pathological tissue image capturing method, and a pathological tissue image capturing program capable of solving the above-described problems.

In order to achieve the above object, the pathological tissue imaging system of the present invention comprises:
A pathological image acquisition unit that captures a pathological tissue image;
An output unit for outputting the pathological tissue image;
A weighting unit that detects an ROI from the pathological tissue image and adds a weight to a pixel located in the ROI;
Based on the weight value added by the weighting unit, a range selection unit that selects an enlarged imaging range for imaging from the pathological tissue image,
The pathological image acquisition unit captures a pathological tissue image obtained by enlarging the enlarged imaging range selected by the range selection unit,
The output unit is a pathological tissue image photographing system that outputs a pathological tissue image photographed by enlarging the enlarged photographing range,
A cancer detection unit for detecting a region suspected of cancer from the pathological tissue image;
The pathological image acquisition unit shoots a pathological tissue image at a magnification capable of recognizing a pattern of cell nuclei contained in a specimen tissue on a pathological tissue slide,
The weighting unit adds a value calculated as a density of a region suspected to be cancer detected from the pathological tissue image by the cancer detection unit to each pixel as a weight of the density,
The range selection unit selects a pixel in the pathological tissue image that maximizes the density weight value, and selects a specified range with the selected pixel as the upper left origin as the enlarged imaging range. Features.

In order to achieve the above object, the pathological tissue imaging method of the present invention comprises:
A pathological tissue imaging method performed by a pathological tissue imaging system that captures and outputs a pathological tissue image,
An imaging step of imaging the pathological tissue image;
A weighting step of detecting an ROI from the pathological tissue image and adding a weight to a pixel located in the ROI;
Based on the weight value added in the weighting step, a range selection step for selecting an enlarged imaging range to be enlarged and photographed from the pathological tissue image;
An enlarged photographing step of photographing a pathological tissue image obtained by enlarging the enlarged range selected in the range selecting step;
An output step of outputting a pathological tissue image taken by enlarging the enlarged imaging range;
A cancer detection step of detecting a region suspected of cancer from the pathological tissue image,
In the imaging step, the pathological tissue image is captured at a magnification capable of recognizing a pattern of cell nuclei contained in the specimen tissue on the pathological tissue slide,
In the weighting step, a value obtained by calculating the density of the suspected cancer area detected from the pathological tissue image in the cancer detection step is added to each pixel as a density weight,
In the range selection step, selecting a pixel having a maximum density value in the pathological tissue image, and selecting a specified range having the selected pixel as an upper left origin as the enlarged imaging range. Features.

In order to achieve the above object, the pathological tissue imaging program of the present invention comprises:
To the pathological tissue imaging system that captures and outputs pathological tissue images,
An imaging procedure for imaging the pathological tissue image;
A weighting procedure for detecting an ROI from the pathological tissue image and adding a weight to a pixel located in the ROI;
Based on the weight value added in the weighting procedure, a range selection procedure for selecting an enlarged imaging range to be enlarged and photographed from the pathological tissue image;
An enlarged imaging procedure for imaging a pathological tissue image obtained by enlarging the enlarged range selected in the range selection procedure;
An output procedure for outputting a pathological tissue image obtained by enlarging the enlarged imaging range;
A cancer detection procedure for detecting a region suspected of cancer from the pathological tissue image, and a pathological tissue imaging program for executing
In the photographing procedure, the pathological tissue image is photographed at a magnification capable of recognizing a pattern of cell nuclei contained in a specimen tissue on the pathological tissue slide,
In the weighting procedure, a value calculated from the density of a region suspected to be cancer detected from the pathological tissue image in the cancer detection procedure is added to each pixel as a weight of the density,
In the range selection procedure, selecting a pixel having a maximum density weight value in the pathological tissue image, and selecting a specified range with the selected pixel as an upper left origin as the enlarged imaging range. Features.

  According to the pathological tissue imaging system of the present invention, the weighting unit automatically detects the ROI from the pathological tissue image, adds the weight to the pixel located in the ROI, and adds the weight to each pixel in the range selection unit. Based on the weight value, a range to be enlarged and photographed is selected, and the pathological image acquisition unit captures a pathological tissue image in which the selected range is enlarged.

  Further, these processes are all automatically performed without any pathologist operation.

  Accordingly, it is possible to detect the ROI from the pathological tissue image without selecting the pathologist, select the detected ROI as the enlarged photographing range, and photograph the pathological tissue image obtained by enlarging the selected enlarged photographing range. can get.

It is a block diagram which shows the structure of the pathological tissue imaging | photography system of the 1st Embodiment of this invention. 3 is a flowchart for explaining the operation of the pathological tissue image capturing system shown in FIG. 1. It is a block diagram which shows the structure of the pathological tissue imaging | photography system of the 2nd Embodiment of this invention. FIG. 4 is a flowchart for explaining the operation of the pathological tissue image capturing system shown in FIG. 3. FIG. It is a block diagram which shows the structure of the pathological tissue imaging | photography system of the 3rd Embodiment of this invention. 6 is a flowchart for explaining the operation of the pathological tissue image capturing system shown in FIG. 5. It is a block diagram which shows the structure of the pathological tissue image imaging system of the 4th Embodiment of this invention. It is a flowchart explaining operation | movement of the pathological tissue imaging | photography system shown in FIG. It is a figure explaining the calculation method of the weight of distance by the distance calculation part shown in FIG. It is a block diagram which shows the structure of the pathological tissue imaging | photography system of the 5th Embodiment of this invention. 12 is a flowchart for explaining the operation of the pathological tissue image capturing system shown in FIG. 10.

The best mode for carrying out the present invention will be described below with reference to the drawings.
(First embodiment)
FIG. 1 shows a configuration of a pathological tissue imaging system according to the first embodiment of the present invention.

  As shown in FIG. 1, the pathological tissue imaging system of the present embodiment includes a pathological image acquisition unit 100, an ROI determination unit 110, and an output unit 120.

  The pathological image acquisition unit 100 is a unit that captures a pathological tissue image at an arbitrary magnification. The pathological image acquisition unit 100 can capture an arbitrary range of a pathological tissue slide in a microscope field of view as a pathological tissue image at an arbitrary magnification, and various magnifications (for example, 0.5 times, 1 time, 1.25). 2 times, 2.5 times, 4 times, 5 times, 10 times, 20 times, 40 times, 60 times, 63 times, 100 times, etc.) and a microscope or scanner CCD camera.

  The ROI determination unit 110 includes a weighting unit 111 and a range selection unit 112. The ROI determination unit 110 detects an ROI from the pathological tissue image captured by the pathological image acquisition unit 100, and enlarges and captures the detected ROI. It is a means to select as a range.

  The weighting unit 111 is a unit that detects the ROI from the pathological tissue image captured by the pathological image acquisition unit 100 and adds a weight to the pixel located in the detected ROI.

  The range selection unit 112 is a unit that selects an enlarged imaging range from the pathological tissue image based on the weight value added by the weighting unit 111. In addition, the range selection unit 112 outputs the position of the selected enlarged imaging range to the pathological image acquisition unit 100.

  In response to this output, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range selected by the range selection unit 112.

  The output unit 120 is a unit that outputs the pathological tissue image captured by the pathological image acquisition unit 100.

  Next, the operation of the pathological tissue image capturing system of the present embodiment will be described using the flowchart of FIG.

  First, the pathological image acquisition unit 100 captures a pathological tissue image at an arbitrary magnification suitable for detecting ROI (step S101).

  Next, the weighting unit 111 detects the ROI from the pathological tissue image captured in step S101 (step S102).

  Next, the weighting unit 111 adds a weight to the pixel located in the ROI detected in step S102 (step S103).

  Next, the range selection unit 112 selects an enlarged imaging range based on the weight value of each pixel from the pathological tissue image to which the weight is added in step S103 (step S104). At this time, the size of the range selected as the enlarged shooting range is an arbitrary size.

  Next, the range selection unit 112 determines whether a predetermined number of enlarged shooting ranges have been obtained (step S105).

  The range selection unit 112 repeats step S104 until a predetermined number of enlarged shooting ranges are obtained.

  Next, the range selection unit 112 outputs the positions of all the enlarged imaging ranges selected in step S105 to the pathological image acquisition unit 100.

  Next, the pathological image acquisition unit 100 captures a pathological group image in which the enlarged imaging range is enlarged with respect to all the enlarged imaging ranges selected in step S105 (step S106). The magnification at the time of photographing a pathological tissue image in which the enlarged photographing range is enlarged is an arbitrary magnification larger than the magnification at which the pathological tissue image is photographed in step S101.

  After that, the output unit 120 outputs all the pathological tissue images that are captured by enlarging the enlarged imaging range in step S106 (step S107).

  As described above, in the pathological tissue imaging system of the present embodiment, the weighting unit 111 detects the ROI from the pathological tissue image, adds a weight to the pixel located in the detected ROI, and the range selection unit 112 The enlarged imaging range is selected based on the weight value added to each pixel, and the pathological image acquisition unit captures a pathological tissue image in which the selected enlarged imaging range is enlarged.

  Further, these processes are all automatically performed without any pathologist operation.

Thereby, the ROI can be automatically detected from the pathological tissue image, the detected ROI can be selected as the enlarged imaging range, and the pathological tissue image obtained by enlarging the selected enlarged imaging range can be obtained.
(Second Embodiment)
FIG. 3 shows a configuration of a pathological tissue image photographing system according to the second embodiment of the present invention.

  As shown in FIG. 3, the pathological tissue imaging system of the present embodiment is different from the first embodiment shown in FIG. 1 in that the ROI determination unit 110 is changed to the ROI determination unit 210. Since other components are the same as those in the first embodiment, the same reference numerals as those in FIG.

  The ROI determination unit 210 includes a weighting unit 211 and a range selection unit 212. The ROI determination unit 210 detects the ROI from the pathological tissue image captured by the pathological image acquisition unit 100, and selects the detected ROI as an enlarged imaging range. It is.

  The weighting unit 211 is a unit that adds a value obtained by calculating the density of cell nuclei in the pathological tissue image captured by the pathological image acquisition unit 100 to each pixel as a weight. In this embodiment, since it is necessary for the weighting unit 211 to calculate the density of cell nuclei contained in the sample tissue on the pathological tissue slide, the pathological image acquisition unit 100 determines that the entire sample tissue on the pathological tissue slide is It is assumed that the pathological tissue image is taken at a magnification that fits in the microscope field of view.

  In order to calculate the density of cell nuclei in the pathological tissue image, the weighting unit 211 detects the edge of the cell nucleus from the pathological tissue image, and sets a region including many detected edges as a region having a high density of cell nuclei. Since the characteristics of cancer are likely to appear in the cell nucleus, the pathologist pays attention to the area where the density of the cell nucleus is high at the time of the histopathological diagnosis of cancer. For this reason, in this embodiment, a region having a high cell nucleus density including many cell nucleus edges is defined as an ROI.

  A specific method for calculating the density of cell nuclei in the pathological tissue image in the weighting unit 211 will be described below. The weighting unit 211 first performs gray scale processing on the pathological tissue image based on the cyan value, then performs smoothing processing, then performs edge detection processing, and detects detected edges. A gradation image is created from the pathological tissue image based on the value. Next, the weighting unit 211 calculates the average value of the edge values of the pixels included in the specified range centered on each pixel of the created gradation image, and calculates the average value of the calculated edge values for each pixel. The density of cell nuclei. Thereafter, the weighting unit 211 adds the calculated value of the density of cell nuclei to each pixel as a weight.

  The range selection unit 212 selects a pixel having the maximum weight value added by the weighting unit 211 in the pathological tissue image, and selects a specified range centered on the selected pixel as an enlarged imaging range. It is. In addition, the range selection unit 212 outputs the position of the selected enlarged photographing range to the pathological image acquisition unit 100.

  In response to this output, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range selected by the range selection unit 212.

  Next, the operation of the pathological tissue image capturing system of the present embodiment will be described using the flowchart of FIG.

  First, the pathological image acquisition unit 100 captures a pathological tissue image at a magnification that allows the entire specimen tissue on the pathological tissue slide to be accommodated in the microscope visual field (step S201).

  Next, the weighting unit 211 grayscales the pathological tissue image based on the cyan value of each pixel of the pathological tissue image in order to calculate the density of cell nuclei in the pathological tissue image captured in step S201. Processing is performed (step S202).

By HE staining, cell nuclei contained in the specimen tissue on the pathological tissue slide are stained blue-purple, and other cytoplasms, fibers, erythrocytes, etc. are stained light red. For this reason, the edge of the cell nucleus, which is the calculation target of the density, can be emphasized by making the pathological tissue image gray scale based on the cyan value. When the pathological tissue image photographed in step S201 is an RGB 256-bit image, the cyan value C can be obtained from Equation 1 below.
(Formula 1)
C = 255-R
Here, the cyan value C can be represented by B + G obtained by mixing B and G in RGB representation (R = Red, G = Green, B = Blue), and R is R in RGB representation.

  Next, the weighting unit 211 performs a smoothing process on the pathological tissue image gray-scaled in step S202 by using a 3 × 3 pixel median filter in order to remove noise (step S203).

  The 3 × 3 median filter is a filter that outputs a median value among 3 × 3 = 9 pixels around the pixel of interest. The median filter may target any number of pixels other than 3 × 3 pixels. Further, an arbitrary smoothing filter may be used without using the median filter.

  Next, the weighting unit 111 calculates a pre-witt edge in order to detect the edge of the cell nucleus in the pathological tissue image smoothed to remove noise in step S203 (step S204).

The pre-wit edge is calculated by calculating the sum of absolute values of the horizontal pre-wit filter f _x (x, y) and the vertical pre-wit filter f _y (x, y) as shown in Equation 2 below. can do. Other filters that detect edges may be used without using the pre-wit filter.
(Formula 2)
Δ (x, y) = | f _x (x, y) | + | f _y (x, y) |
Next, the weighting unit 211 obtains the minimum value and the maximum value of the Prewitt filter absolute value sum, which is the edge value of each pixel obtained in step S204.

  Next, the weighting unit 211 divides the difference between the maximum value and the minimum value of the obtained Prewitt filter absolute value sum into five equal parts, and distributes each pixel into five groups according to the Prewitt filter absolute value sum value. At this time, each pixel is added with the assigned group value (for example, 1 to 5) as a weight. Here, it is assumed that a group having a larger value of the sum of absolute values of the Prewitt filter of the pixel to be distributed has a larger group value. Further, the method of dividing the difference between the maximum value and the minimum value of the absolute value sum of the Prewitt filters of each pixel is not limited to 5 and may be any division method and any number of divisions.

  Next, the weighting unit 211 creates a five-tone image from the pathological tissue image based on the weight value added to each pixel (step S205).

  Next, the weighting unit 211 calculates an average value of the weight values of the pixels included in the specified range centered on each pixel with respect to the 5-tone image created in step S205, and creates an edge average image. (Step S206). The average value of the weight values calculated at this time is used as the final weight value of each pixel. As a result, a larger weight is added to a pixel located in a region including many cell nucleus edges around the periphery, that is, a region having a high density of cell nuclei.

  Here, the size of the prescribed range centered on each pixel used for calculating the average value of the weight around each pixel is a range that is 33 pixels away from the center pixel. The size of the specified range used to calculate the average value of the surrounding weights may be a range that is an arbitrary number of pixels away from the center pixel.

  Next, the range selection unit 212 expands the range around the pixel having the maximum weight value, that is, the pixel located in the region with the highest density of cell nuclei, from the edge average image created in step S206. (Step S207). At this time, the size of the range selected as the enlarged shooting range is an arbitrary size.

  Next, the range selection unit 212 sets the weight value of the pixels included in the enlarged shooting range selected in step S207 to 0 so that the same range is not selected as the enlarged shooting range multiple times (step S208).

  Next, the range selection unit 212 determines whether a predetermined number of enlarged shooting ranges have been obtained (step S209).

  The range selection unit 212 repeats steps S207 and S208 until a predetermined number of enlarged shooting ranges are obtained.

  Next, the range selection unit 212 outputs the positions of all the enlarged shooting ranges selected in step S209 to the pathological image acquisition unit 100.

  Next, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range with respect to all the enlarged imaging ranges selected in step S209 (step S210). The magnification at the time of photographing the pathological tissue image in which the magnified photographing range is enlarged is an arbitrary magnification larger than the magnification at which the pathological tissue image is photographed in step S201.

  Thereafter, the output unit 120 outputs all the pathological tissue images that have been captured by enlarging the enlarged imaging range in step S210 (step S211).

  As described above, in the pathological tissue imaging system of the present embodiment, the weighting unit 211 detects a region having a high density of cell nuclei from which a characteristic of cancer is likely to appear from the pathological tissue image as an ROI, and is positioned in the ROI. A pathology obtained by adding a weight to a pixel to be processed, selecting an enlarged imaging range based on the weight value added to each pixel in the range selection unit 212, and expanding the selected enlarged imaging range in the pathological image acquisition unit 100 It is configured to take tissue images.

  Further, these processes are all automatically performed without any pathologist operation.

Thereby, the ROI can be automatically detected from the pathological tissue image, the detected ROI can be selected as the enlarged imaging range, and the pathological tissue image obtained by enlarging the selected enlarged imaging range can be obtained.
(Third embodiment)
FIG. 5 shows a configuration of a pathological tissue image photographing system according to the third embodiment of the present invention.

  As shown in FIG. 5, the pathological tissue imaging system according to the present embodiment is different from the second embodiment shown in FIG. 3 in that the ROI determination unit 210 is changed to the ROI determination unit 310, and the cancer The difference is that a detection unit 330 is added. Since other components are the same as those in the second embodiment, the same reference numerals as those in FIG.

  The cancer detection unit 330 is means for detecting a region suspected of being cancerous from the pathological tissue image captured by the pathological image acquisition unit 100. In addition, the cancer detection unit 330 adds a mark to each pixel located in a region suspected of being detected from a pathological tissue image. Here, as a technique for detecting a region suspected of being cancerous, a technique that quantitatively represents the features of a pathological tissue image described in Patent Document 1 is used. Any other learning method or image processing method may be used as a method of detecting a region suspected of having cancer.

  In the cancer detection unit 330, the magnification for capturing the pathological tissue image is too small when the entire specimen tissue on the pathological tissue slide used in the second embodiment fits in the microscope field of view. Therefore, the area suspected of being cancerous cannot be detected. In order for the cancer detection unit 330 to detect a region suspected of being cancerous, it is necessary to be able to recognize the pattern of cell nuclei in the pathological tissue image. Therefore, in the present embodiment, the pathological image acquisition unit 100 has a magnification that allows the cancer detection unit 330 to detect a region suspected of cancer, and a pattern of cell nuclei included in the sample tissue on the pathological tissue slide. It is assumed that a pathological tissue image is taken at a magnification that can be recognized by the user. In general, the magnification at which a cell nucleus pattern can be recognized in a microscopic field (in a pathological tissue image) is about 10 times.

  The ROI determination unit 310 includes a weighting unit 311 and a range selection unit 312. The region that is suspected to be cancer detected from the pathological tissue image by the cancer detection unit 330 is defined as ROI, and the ROI is enlarged imaging range. It is a means to select as.

  The weighting unit 311 is a unit that adds a value obtained by calculating the density of a region suspected of cancer detected from the pathological tissue image by the cancer detection unit 330 to each pixel as a density C. The weighting unit 311 adds a constant value to the pixels located at the upper left origin of the specified range for all patterns in the specified range including the pixels to which the cancer detection unit 330 has added marks in the pathological tissue image. The obtained value is set as a density C.

  In the present embodiment, a region suspected of being cancer detected by the cancer detection unit 330 is defined as ROI.

  The range selection unit 312 selects a pixel in the pathological tissue image in which the value of the density weight C added by the weighting unit 311 is maximized, and enlarges the specified range with the selected pixel as the upper left origin. It is a means to select as a range. The range selection unit 312 outputs the position of the selected enlarged imaging range to the pathological image acquisition unit 100.

  Upon receiving this output, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range selected by the range selection unit 312.

  Next, the operation of the pathological tissue imaging system of the third embodiment will be described using the flowchart of FIG.

  First, the pathological image acquisition unit 100 captures a pathological tissue image at a magnification capable of recognizing the pattern of the cell nucleus of the specimen tissue on the pathological tissue slide (step S301).

  Next, the cancer detection unit 330 uses the technique of quantitatively expressing the characteristics of the pathological tissue image described in Patent Document 1 from the pathological tissue image photographed in step S301 to identify a region suspected of having cancer. It detects and adds a mark to the pixel located in the detected area (step S302).

  Next, in the weighting unit 311, the cancer detection unit 330 adds marks to all the pixels of the pathological tissue image in which marks are added to the pixels located in the region suspected of having cancer in step S <b> 302. It is determined whether it is a pixel (step S303).

  In step S303, when a mark is added to the target pixel as the determination target, the weighting unit 311 uses the upper left as the origin for all the patterns in the specified range of the width W and the height H including the target pixel. Then, a certain value is added to the pixel located at the upper left origin, and the added value is set as the density weight C (step S304). The congestion weight C is stored in each pixel and added. For this reason, the value of the density weight C increases as the number of pixels to which the mark is added is within the prescribed range of the width W and the height H with the self as the upper left origin. Therefore, a pixel having a higher density weight C means a pixel located in an area where the density of the suspected cancer area is higher, that is, an area having high importance as an ROI. Here, 1 is added to the value of the density weight C in one addition, but it may be an arbitrary value instead of 1.

  Further, the values of the width W and the height H, which are the size of the specified range including the target pixel, are arbitrary values.

  In addition, here, the method of adding weight to the pixel located at the upper left origin is shown, but in addition, the method of adding weight using the upper right, lower left and lower right as the origin, the method of adding weight to the center pixel, etc. A method of adding arbitrary weights may be used.

  Next, the weighting unit 311 determines whether the processing in step S303 and step S304 has been performed on all the pixels of the pathological tissue image (step S305).

  The weighting unit 311 repeats the processes of Step S303 and Step S304 until all the pixels of the pathological tissue image are performed (Step S306).

  Next, the range selection unit 312 selects a pixel having the maximum value of the density weight C added in step S304, and sets a predetermined range of width W and height H with the selected pixel as the upper left origin. The enlarged photographing range is selected (step S307).

  Next, the range selection unit 312 selects the selected enlarged shooting range from the density weight C of each pixel included in the enlarged shooting range selected in step S307 so as not to select the same range as the enlarged shooting range multiple times. The number of pixels to which the mark included in is added is subtracted (step S308). For example, if there are three pixels with a mark in the selected enlarged photographing range in step S307, 3 is subtracted from the value C of the density C of all the pixels in the enlarged photographing range. Also, the weight subtraction method is not limited to this, and a method of setting all the density weights C of each pixel in the enlarged photographing range to 0 or a method of subtracting other arbitrary weights may be used. .

  Next, the range selection unit 312 determines whether or not a predetermined number of enlarged shooting ranges have been obtained (step S309).

  The range selection unit 312 repeats step S307 and step S308 until a predetermined number of enlarged shooting ranges are obtained.

  Next, the range selection unit 312 outputs the positions of all the enlarged shooting ranges selected in step S309 to the pathological image acquisition unit 100.

  Next, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range with respect to all the enlarged imaging ranges selected in step S309 (step S310). The magnification at the time of photographing the pathological tissue image in which the magnified photographing range is enlarged is an arbitrary magnification larger than the magnification at which the pathological tissue image is photographed in step S301.

  After that, the output unit 120 outputs all the pathological tissue images that are captured by enlarging the enlarged imaging range in step S310 (step S311).

  As described above, in the pathological tissue imaging system of the present embodiment, the cancer detection unit 330 detects a region suspected of cancer from the pathological tissue image, and the weighting unit 311 detects the detected cancer. The suspicious area is defined as ROI, and a weight is added to the pixel located in the ROI. The range selection unit 312 selects an enlarged imaging range based on the weight value added to each pixel, and the pathological image acquisition unit In this configuration, a pathological tissue image obtained by enlarging the selected enlarged range photographing range is photographed.

  Further, these processes are all automatically performed without any pathologist operation.

Thereby, the ROI can be automatically detected from the pathological tissue image, the detected ROI can be selected as the enlarged imaging range, and the pathological tissue image obtained by enlarging the selected enlarged imaging range can be obtained.
(Fourth embodiment)
FIG. 7 shows a configuration of a pathological tissue image photographing system according to the fourth embodiment of the present invention.

  As shown in FIG. 7, the pathological tissue imaging system of the present embodiment is different from the third embodiment shown in FIG. 5 in that the ROI determination unit 310 is changed to the ROI determination unit 410. The ROI determination unit 410 is different from the ROI determination unit 310 in that the range selection unit 312 is changed to the range selection unit 412 and a distance calculation unit 413 is added. Since other components are the same as those in the third embodiment, the same reference numerals as those in FIG.

  The ROI determination unit 410 includes a weighting unit 311, a range selection unit 412, and a distance calculation unit 413, and an area suspected of cancer detected from a pathological tissue image by the cancer detection unit 330 is defined as ROI. , ROI is a means for selecting an enlarged photographing range.

  The distance calculation unit 413 calculates a value obtained by calculating the distance between each pixel and the center pixel of the specified range for all patterns in the specified range including each pixel in the pathological tissue image captured by the pathological image acquisition unit 100. This is a means for adding a weight D to the pixel located at the upper left origin of the specified range.

  The distance weight D represents how far the target pixel of interest is away from the center pixel in the specified range considered now. For this reason, the ROI is present at a position closer to the center in a prescribed range having a pixel having a smaller distance weight D at the upper left origin. Accordingly, when the range selection unit 412 selects the enlarged imaging range, the ROI is held at the center of the pathological tissue image by selecting a specified range in which the value of the distance weight D of the pixel located at the upper left origin is small. I can go.

  The range selection unit 412 selects a pixel having the maximum value of the density weight C added by the weighting unit 311 in the pathological tissue image, and includes a plurality of pixels having the maximum value of the density weight C. If there is a pixel, the pixel having the smallest distance weight D added by the distance calculator 413 is selected from these pixels, and the specified range with the selected pixel as the upper left origin is selected as the enlarged photographing range. It is means to do. In addition, the range selection unit 412 outputs the position of the selected enlarged photographing range to the pathological image acquisition unit 100.

  In response to this output, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range selected by the range selection unit 412.

  Next, the operation of the pathological tissue imaging system of this embodiment will be described using the flowchart of FIG.

  First, the pathological image acquisition unit 100 captures a pathological tissue image at a magnification capable of recognizing the pattern of the cell nucleus of the specimen tissue on the pathological tissue slide (step S401).

  Next, the cancer detection unit 330 uses the technique of quantitatively expressing the characteristics of the pathological tissue image described in Patent Document 1 from the pathological tissue image photographed in step S401 to identify a region suspected of having cancer. Detection is performed, and a mark is added to the pixel located in the detected area (step S402).

  Next, in the weighting unit 311, the cancer detection unit 330 adds a mark to all pixels of the pathological tissue image in which the mark is added to the pixel located in the region suspected of having cancer in step S <b> 402. It is determined whether it is a pixel (step S403).

  In step S403, when a mark is added to the target pixel as a determination target, the weighting unit 311 uses the upper left as the origin for all the patterns in the specified range of the width W and the height H including the target pixel. Then, a certain value is added to the pixel located at the upper left origin, and the added value is set as the density weight C (step S404).

  Next, the distance calculation unit 413 adds a distance weight D to each pixel to the pathological tissue image to which the congestion weight C is added in step S404 (step S405).

  The distance weight D is calculated as follows. The distance calculation unit 413 includes all pixels in the prescribed range of the width W and the height H with the origin at the upper left, including pixels located in a region suspected of being cancer that has been marked by the cancer detection unit 330. On the other hand, the distance between the target pixel and the center pixel (X = W / 2, Y = H / 2) of the specified range that is currently considered is calculated, and the pixel located at the upper left origin of the specified range that is currently considered is calculated. It is added as a distance weight D.

  For example, as shown in FIG. 9, a range of width W = 7 and height H = 7 is considered as a specified range. If the target pixel to which the mark is added by the cancer detection unit 330 is at the coordinate position (2, 2), the target pixel has a width W = 7 and a height H = 7 within the specified range. This means that the distance is 2√2 from the center pixel (4, 4). In this case, a distance weight D = 2√2 is added to the pixel (1, 1) located at the upper left origin of the specified range considered now. When the distance weight D is already added to the pixel (1, 1) located at the upper left origin, the smaller value is adopted as the distance weight D.

  As described above, here, in step S405, the distance between the target pixel located in the region suspected of being marked with cancer and the central pixel in the prescribed range of width W and height H under consideration. Although the distance weight D is calculated, a distance calculation map may be created in advance, and the distance between the corresponding target pixel and the center pixel may be acquired from the distance calculation map in step S405.

  In addition, here, a method of adding a distance weight D to a pixel located at the upper left origin with the upper left of the specified range being considered as the origin is shown, but in addition, a distance with the upper right, lower left, lower right, etc. as the origin Alternatively, a method of adding the weight D, a method of adding the distance weight D to the center pixel, or a method of adding any other weight may be used.

  In addition, when the distance weight D has already been added to the pixel at the upper left origin, the smaller value is adopted as the distance weight D, but a method of adding them or other methods may be used.

  Note that the values of the width W and the height H, which are the size of a specified range including the target pixel, are arbitrary values.

  Next, the weighting unit 311 and the distance calculation unit 413 determine whether or not the processing in steps S403 to S405 has been performed on all the pixels of the pathological tissue image (step S406).

  The weighting unit 311 and the distance calculation unit 413 repeat the processing in steps S403 to S405 for all the pixels of the pathological tissue image (step S407).

  Next, the range selection unit 412 determines whether there are a plurality of pixels in the pathological tissue image that have the maximum value of the density weight C added in step S404 (step S408).

  The range selection unit 412 selects the pixel having the maximum density weight C value when there are not a plurality of pixels having the maximum density weight C value, and uses the selected pixel as the upper left origin. Then, the specified range of the height H is selected as the enlarged photographing range (step S409).

  When there are a plurality of pixels having the maximum density weight C, the range selection unit 412 selects a pixel having the minimum distance weight D among the pixels, and selects the selected pixel in the upper left. A prescribed range of width W and height H as the origin is selected as the enlarged photographing range (step S410).

  Next, the range selection unit 412 selects the selected enlargement from the density weight C of each pixel in the enlarged shooting range selected in steps S409 to S410 so as not to select the same range as the enlarged shooting range a plurality of times. The number of pixels to which marks included in the shooting range are added is subtracted (step S411).

  Next, the range selection unit 412 determines whether a predetermined number of enlarged shooting ranges have been obtained (step S412).

  The range selection unit 412 repeats step S408 to step S411 until a predetermined number of enlarged shooting ranges are obtained.

  Next, the range selection unit 412 outputs the positions of all selected enlarged imaging ranges to the pathological image acquisition unit 100.

  Next, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range with respect to all the enlarged imaging ranges selected in Step S412 (Step S413). The magnification at the time of photographing the pathological tissue image in which the magnified photographing range is enlarged is an arbitrary magnification larger than the magnification at which the pathological tissue image is photographed in step S401.

  Thereafter, the output unit 120 outputs all the pathological tissue images that have been captured by enlarging the enlarged imaging range in step S413 (step S414).

  In the present embodiment, the distance weight D is added in step S405. However, after step S408, only when there are a plurality of pixels having the maximum value of the density weight C, the distance for these pixels is determined. May be calculated.

  As described above, in the pathological tissue imaging system of the present embodiment, the cancer detection unit 330 detects a region suspected of cancer from the pathological tissue image, and the weighting unit 311 and the distance calculation unit 413 detect them. The region suspected of being cancer is set as ROI, and a weight is added to the pixel located in the ROI, and the range selection unit 412 selects an enlarged photographing range based on the value of the weight added to each pixel, The pathological image acquisition unit 100 is configured to capture a pathological tissue image in which the selected enlarged imaging range is enlarged.

  Further, these processes are all automatically performed without any pathologist operation.

Thereby, the ROI can be automatically detected from the pathological tissue image, the detected ROI can be selected as the enlarged imaging range, and the pathological tissue image obtained by enlarging the selected enlarged imaging range can be obtained.
(Fifth embodiment)
FIG. 10 shows a configuration of a pathological tissue image photographing system according to the fifth embodiment of the present invention.

  As shown in FIG. 10, the pathological tissue imaging system of the present embodiment is different from the fourth embodiment shown in FIG. 7 in that the ROI determination unit 410 is changed to the ROI determination unit 510. The ROI determination unit 510 differs from the ROI determination unit 410 in that the range selection unit 412 is changed to the range selection unit 512 and a tissue region calculation unit 514 is added. Since other components are the same as those in the fourth embodiment, the same reference numerals as those in FIG.

  The ROI determination unit 510 includes a weighting unit 311, a range selection unit 512, a distance calculation unit 413, and a tissue region calculation unit 514. The cancer detected by the cancer detection unit 330 from the pathological tissue image This is a means for selecting a region suspected of being ROI and selecting ROI as an enlarged photographing range.

  The tissue region calculation unit 514 detects a tissue region from the pathological tissue image captured by the pathological image acquisition unit 100, and calculates the number of pixels located in the tissue region within a specified range with each pixel as the upper left origin. A means for adding a value to each pixel as a weight R of the tissue region. The tissue region calculation unit 514 detects pixels that satisfy all of the following three conditions from the pathological tissue image, and sets these pixels as pixels located in the tissue region.

  Conditions when the tissue region calculation unit 514 detects pixels located in the tissue region are as follows.

  (Condition 1) The maximum value of the luminances of R, G, and B is 64 or more.

  (Condition 2) The value obtained by subtracting the luminance of G from the smaller one of the luminances of R and B is 10 or more.

  (Condition 3) A value obtained by multiplying the difference between the maximum value and the minimum value among the luminances of R, G, and B by 255 and dividing the value by the maximum value among the luminances of R, G, and B is 32 or more. It becomes.

  The range selection unit 512 selects a pixel in the pathological tissue image that has the maximum density weight C value added by the weighting unit 311 and has a plurality of pixels that have the maximum density weight C value. If there is a pixel, the pixel having the smallest distance weight D value added by the distance calculation unit 413 is selected from these pixels, and there are a plurality of pixels having the smallest distance weight D value. From these pixels, a pixel having the maximum value of the tissue region weight R added by the tissue region calculation unit 514 is selected, and a specified range with the selected pixel as the upper left origin is selected as the enlarged photographing range. Means. The range selection unit 512 outputs the position of the selected enlarged imaging range to the pathological image imaging unit 100.

  In response to this output, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range selected by the range selection unit 512.

  Next, the operation of the pathological tissue imaging system of this embodiment will be described using the flowchart of FIG.

  First, the pathological image acquisition unit 100 acquires a pathological tissue image at a magnification capable of recognizing the pattern of the cell nucleus of the specimen tissue on the pathological tissue slide (step S501).

  Next, the cancer detection unit 330 uses the technique of quantitatively expressing the characteristics of the pathological tissue image described in Patent Document 1 from the pathological tissue image photographed in step S501 to identify an area suspected of having cancer. Detection is performed and a mark is added to the pixel located in the detected area (step S502).

  Next, in the weighting unit 311, the cancer detection unit 330 adds the mark to all the pixels of the pathological tissue image in which the mark is added to the pixel located in the region suspected of having cancer in step S <b> 502. It is determined whether it is a pixel (step S503).

  In step S503, when a mark is added to the target pixel as a determination target, the weighting unit 311 uses the upper left as the origin for all the patterns in the specified range of the width W and the height H including the target pixel. Then, a certain value is added to the pixel located at the upper left origin, and the added value is set as a density weight C (step S504).

  Next, the distance calculation unit 413 adds a distance weight D to each pixel to the pathological tissue image to which the congestion weight C is added in step S504 (step S505).

  Next, the weighting unit 311 and the distance calculation unit 413 determine whether or not the processing in steps S503 to S505 has been performed on all the pixels of the pathological tissue image (step S506).

  The weighting unit 311 and the distance calculation unit 413 repeat the processing in steps S503 to S505 for all the pixels of the pathological tissue image (step S507).

  Next, the range selection unit 512 determines whether there are a plurality of pixels in the pathological tissue image that have the maximum value of the density weight C added in step S504 (step S508).

  When there are not a plurality of pixels having the maximum density weight C value, the range selection unit 512 selects the pixel having the maximum density weight C value, and uses the selected pixel as the upper left origin. Then, the specified range of the height H is selected as the enlarged photographing range (step S509).

  When there are a plurality of pixels having the maximum density weight C in step S508, the range selection unit 512 determines whether there are a plurality of pixels having the minimum distance weight D from among those pixels. Determination is made (step S510).

  When there are not a plurality of pixels having the smallest distance weight D value, the range selection unit 512 selects the pixel having the smallest distance weight D value, and uses the selected pixel as the upper left origin for the width W and the high The specified range of the height H is selected as the enlarged photographing range (step S511).

  When there are a plurality of pixels having the smallest distance weight D value in step S510, the tissue region calculation unit 514 has a predetermined range of width W and height H with the target pixel as the upper left origin. Calculate how many pixels located in the tissue area that satisfy all three conditions described above are included, and add the calculated number of pixels located in the tissue area to the target pixel as the weight R of the tissue area (Step S512).

  At this time, the tissue region calculation unit 514 selects a pixel having a maximum value of 64 or more among the luminances of R, G, and B in the pathological tissue image in order to detect pixels located in the tissue region, From the selected pixels, a pixel in which the value obtained by subtracting the luminance of G from the smaller one of the luminances of R and B is 10 or more is selected, and among the luminances of R, G, and B, among the selected pixels A pixel obtained by multiplying the value obtained by multiplying the value between the maximum value and the minimum value by 255 by the maximum value among the luminances of R, G, and B is selected to be 32 or more, and the selected pixel is selected from the tissue. It is detected as a pixel located in the area.

  Here, pixels satisfying all the above three conditions are detected as pixels located in the tissue region, but other tissue region detection methods may be used.

  Here, the tissue region is detected from the pathological tissue image in step S512. However, a method may be used in which a mask image in which the tissue region is detected is created in advance and the mask image is applied in step S512.

  Further, after step S505, the weight R of the tissue region may be calculated and added to each pixel in advance.

  Next, when there are a plurality of pixels having the smallest distance weight value D in step S510, the range selection unit 512 selects the pixel having the largest tissue region weight value R added in step S512. Then, a predetermined range of width W and height H with the selected pixel as the upper left origin is selected as the enlarged photographing range (step S513).

  Next, the range selection unit 512 selects the enlargement selected from the density weight C of each pixel in the enlarged shooting range selected in steps S509 to S513 so as not to select the same range as the enlarged shooting range multiple times. The number of pixels to which marks included in the shooting range are added is subtracted (step S514).

  Next, the range selection unit 512 determines whether or not a predetermined number of enlarged shooting ranges have been obtained (step S515).

  The range selection unit 512 repeats Steps S406 to S413 until a predetermined number of enlarged shooting ranges are obtained.

  Next, the range selection unit 512 outputs the positions of all the enlarged imaging ranges selected in step S515 to the pathological image acquisition unit 100.

  Next, the pathological image acquisition unit 100 captures a pathological tissue image obtained by enlarging the enlarged imaging range with respect to all the enlarged imaging ranges selected in Step S515 (Step S516). The magnification at the time of photographing the pathological tissue image in which the magnified photographing range is enlarged is an arbitrary magnification larger than the magnification at which the pathological tissue image is photographed in step S501.

  Thereafter, the output unit 120 outputs all the pathological tissue images that have been captured by enlarging the enlarged imaging range in step S516 (step S517).

  As described above, in the pathological tissue imaging system of the present embodiment, the cancer detection unit 330 detects a region suspected of cancer from the pathological tissue image, and the weighting unit 311, the distance calculation unit 413, and the tissue region. In the calculation unit 514, an area suspected of being detected as cancer is defined as ROI, and a weight is added to a pixel located in the ROI. In the range selection unit 512, an enlargement is performed based on the weight value added to each pixel. The imaging range is selected, and the pathological image acquisition unit 100 captures a pathological tissue image in which the selected enlarged imaging range is enlarged.

  Further, these processes are all automatically performed without any pathologist operation.

  Thereby, the ROI can be automatically detected from the pathological tissue image, the detected ROI can be selected as the enlarged imaging range, and the pathological tissue image obtained by enlarging the selected enlarged imaging range can be obtained.

  Note that the above-described pathological tissue imaging system of the first to fifth embodiments is an example, and the configuration and operation thereof can be appropriately changed without departing from the gist of the invention.

  For example, first, a region having a high density of cell nuclei is detected using the second embodiment, a pathological tissue image obtained by enlarging the detected region having a high density of cell nuclei is captured, and then the second embodiment is changed to the second embodiment. Based on the pathological tissue image photographed using, the region suspected of cancer is detected using the third to fifth embodiments, the detected region is further enlarged and selected as a region to be photographed, and the selected region An enlarged pathological tissue image may be taken.

  In addition, for example, in the third to fifth embodiments, the cancer detection unit 330 detects a suspicious area of cancer, but by incorporating a system for detecting other diseases, pathologies other than cancer You may enable it to respond to a tissue diagnosis.

  Further, for example, by combining with the histopathological diagnosis support system described in Patent Document 1, the pathological tissue image obtained by detecting the ROI from the pathological tissue image and enlarging the detected ROI without the operation of the pathologist It may be an automatic histopathological diagnosis support system that performs accurate histopathological diagnosis based on the above.

  Note that the pathological tissue imaging system of the present invention is a program that realizes its function on a recording medium that can be read by the pathological tissue imaging system in addition to that realized by dedicated hardware as described above. The program recorded and recorded on this recording medium may be read by the pathological tissue imaging system and executed. The recording medium readable by the pathological tissue imaging system refers to a recording medium such as a flexible disk, a magneto-optical disk, a CD-ROM, or a storage device such as a hard disk device built in the pathological tissue imaging system. Furthermore, the recording medium that can be read by the histopathological imaging system is constant like a volatile memory inside the histopathological imaging system for a short time, as in the case of transmitting a program via the Internet. Includes those holding time programs.

100 Pathological image acquisition unit 110, 210, 310, 410, 510 ROI determination unit 111, 211, 311 Weighting unit 112, 212, 312, 412, 512 Range selection unit 120 Output unit 330 Cancer detection unit 413 Distance calculation unit 514 Organizational area calculator

Claims (15)

A pathological image acquisition unit that captures a pathological tissue image;
An output unit for outputting the pathological tissue image;
A weighting unit that detects an ROI from the pathological tissue image and adds a weight to a pixel located in the ROI;
Based on the weight value added by the weighting unit, a range selection unit that selects an enlarged imaging range for imaging from the pathological tissue image,
The pathological image acquisition unit captures a pathological tissue image obtained by enlarging the enlarged imaging range selected by the range selection unit,
The output unit is a pathological tissue image photographing system that outputs a pathological tissue image photographed by enlarging the enlarged photographing range,
A cancer detection unit for detecting a region suspected of cancer from the pathological tissue image;
The pathological image acquisition unit shoots a pathological tissue image at a magnification capable of recognizing a pattern of cell nuclei contained in a specimen tissue on a pathological tissue slide,
The weighting unit adds a value calculated as a density of a region suspected to be cancer detected from the pathological tissue image by the cancer detection unit to each pixel as a weight of the density,
The range selection unit selects a pixel in the pathological tissue image that maximizes the density weight value, and selects a specified range with the selected pixel as the upper left origin as the enlarged imaging range. A featured pathological tissue imaging system. The cancer detection unit adds a mark to each pixel located in a region suspected of being cancer detected from the pathological tissue image,
The weighting unit is fixed to pixels located at the upper left origin of the specified range for all patterns in the specified range including pixels to which the cancer detection unit has added marks in the pathological tissue image. The pathological tissue imaging system according to claim 1, wherein a value obtained by adding the values is used as a density weight. A pixel located at the upper left origin of the specified range with a value obtained by calculating the distance between each pixel and the center pixel of the specified range for all patterns in the specified range including each pixel in the pathological tissue image A distance calculation unit for adding the weight as a distance weight to
The range selection unit selects a pixel having the maximum density weight value in the pathological tissue image, and when there are a plurality of the pixels, the distance weight value is the smallest among the pixels. The pathological tissue imaging system according to claim 2, wherein the predetermined range is selected as the enlarged imaging range with the selected pixel as the upper left origin. A tissue region is detected from the pathological tissue image, and a value obtained by calculating the number of pixels located in the tissue region included in the specified range with each pixel as an upper left origin is added to each pixel as a weight of the tissue region. It further has an organization area calculation unit,
The range selection unit selects a pixel in the pathological tissue image that has the maximum density weight value, and when there are a plurality of the pixels, the distance weight value is the smallest among the pixels. If there are a plurality of such pixels, the pixel having the maximum weight value of the tissue region is selected from the pixels, and the specified range with the selected pixel as the upper left origin is The pathological tissue image photographing system according to claim 3, which is selected as an enlarged photographing range.   The tissue region calculation unit selects a pixel having a maximum value of 64 or more among the luminances of R, G, and B in the pathological tissue image, and selects the smaller one of the luminances of R and B from the selected pixels. A pixel in which the value obtained by subtracting the luminance of G from the value of G is 10 or more is selected, and 255 is selected as the value obtained by taking the difference between the maximum value and the minimum value among the luminances of R, G, and B from the selected pixels. 5. The pathological tissue imaging system according to claim 4, wherein a pixel having a value obtained by dividing the multiplied value by the maximum value is 32 or more, and the selected pixel is detected as a pixel located in the tissue region. A pathological tissue imaging method performed by a pathological tissue imaging system that captures and outputs a pathological tissue image,
An imaging step of imaging the pathological tissue image;
A weighting step of detecting an ROI from the pathological tissue image and adding a weight to a pixel located in the ROI;
Based on the weight value added in the weighting step, a range selection step for selecting an enlarged imaging range to be enlarged and photographed from the pathological tissue image;
An enlarged photographing step of photographing a pathological tissue image obtained by enlarging the enlarged range selected in the range selecting step;
An output step of outputting a pathological tissue image taken by enlarging the enlarged imaging range;
A cancer detection step of detecting a region suspected of cancer from the pathological tissue image,
In the imaging step, the pathological tissue image is captured at a magnification capable of recognizing a pattern of cell nuclei contained in the specimen tissue on the pathological tissue slide,
In the weighting step, a value obtained by calculating the density of the suspected cancer area detected from the pathological tissue image in the cancer detection step is added to each pixel as a density weight,
In the range selection step, selecting a pixel having a maximum density value in the pathological tissue image, and selecting a specified range having the selected pixel as an upper left origin as the enlarged imaging range. A feature of a pathological tissue imaging method. In the cancer detection step, a mark is added to each pixel located in a region suspected of being cancer detected from the pathological tissue image,
In the weighting step, for all the patterns in the specified range including the pixels to which the mark was added in the cancer detection step in the pathological tissue image, constant in pixels located at the upper left origin of the specified range The pathological tissue image capturing method according to claim 6, wherein a value obtained by adding the values is used as a density weight. A pixel located at the upper left origin of the specified range with a value obtained by calculating the distance between each pixel and the center pixel of the specified range for all patterns in the specified range including each pixel in the pathological tissue image A distance calculating step for adding to the
In the range selection step, a pixel having the maximum density weight value in the pathological tissue image is selected, and when there are a plurality of the pixels, the distance weight value is the smallest among the pixels. The pathological tissue image photographing method according to claim 7, wherein a pixel to be selected is selected, and the specified range having the selected pixel as an upper left origin is selected as the enlarged photographing range. A tissue region is detected from the pathological tissue image, and a value obtained by calculating the number of pixels located in the tissue region included in the specified range with each pixel as an upper left origin is added to each pixel as a weight of the tissue region. And further comprising a tissue area calculation step,
In the range selection step, a pixel having the maximum density value in the pathological tissue image is selected, and when there are a plurality of the pixels, the distance weight value is the minimum among the pixels. If there are a plurality of such pixels, the pixel having the maximum weight value of the tissue region is selected from the pixels, and the specified range with the selected pixel as the upper left origin is The pathological tissue image photographing method according to claim 8, wherein the pathological tissue image photographing method is selected as an enlarged photographing range.   In the tissue region calculation step, a pixel having a maximum value of 64 or more among the luminances of R, G, and B in the pathological tissue image is selected, and the smaller one of the luminances of R and B from the selected pixels A pixel in which the value obtained by subtracting the luminance of G from the value of G is 10 or more is selected, and 255 is selected as the value obtained by taking the difference between the maximum value and the minimum value among the luminances of R, G, and B from the selected pixels. The pathological tissue imaging method according to claim 9, wherein a pixel in which a value obtained by dividing the multiplied value by the maximum value is 32 or more is selected, and the selected pixel is detected as a pixel located in the tissue region. To the pathological tissue imaging system that captures and outputs pathological tissue images,
An imaging procedure for imaging the pathological tissue image;
A weighting procedure for detecting an ROI from the pathological tissue image and adding a weight to a pixel located in the ROI;
Based on the weight value added in the weighting procedure, a range selection procedure for selecting an enlarged imaging range to be enlarged and photographed from the pathological tissue image;
An enlarged imaging procedure for imaging a pathological tissue image obtained by enlarging the enlarged range selected in the range selection procedure;
An output procedure for outputting a pathological tissue image obtained by enlarging the enlarged imaging range;
A cancer detection procedure for detecting a region suspected of cancer from the pathological tissue image, and a pathological tissue imaging program for executing
In the photographing procedure, the pathological tissue image is photographed at a magnification capable of recognizing a pattern of cell nuclei contained in a specimen tissue on the pathological tissue slide,
In the weighting procedure, a value calculated from the density of a region suspected to be cancer detected from the pathological tissue image in the cancer detection procedure is added to each pixel as a weight of the density,
In the range selection procedure, selecting a pixel having a maximum density weight value in the pathological tissue image, and selecting a specified range with the selected pixel as an upper left origin as the enlarged imaging range. A pathological tissue imaging program characterized. In the cancer detection procedure, a mark is added to each pixel located in a region suspected of being cancer detected from the pathological tissue image,
In the weighting procedure, the pixels located at the upper left origin of the specified range are constant for all patterns in the specified range including the pixels to which the mark was added in the cancer detection procedure in the pathological tissue image. The pathological tissue imaging program according to claim 11, wherein a value obtained by adding the two values is used as a density weight. A pixel located at the upper left origin of the specified range with a value obtained by calculating the distance between each pixel and the center pixel of the specified range for all patterns in the specified range including each pixel in the pathological tissue image A distance calculation procedure for adding a distance weight to
In the range selection procedure, a pixel having the maximum density weight value is selected in the pathological tissue image, and when there are a plurality of the pixels, the distance weight value is the smallest among the pixels. The pathological tissue imaging program according to claim 12, wherein a pixel to be selected is selected, and the specified range with the selected pixel as an upper left origin is selected as the enlarged imaging range. A tissue region is detected from the pathological tissue image, and a value obtained by calculating the number of pixels located in the tissue region included in the specified range with each pixel as an upper left origin is added to each pixel as a weight of the tissue region. Further comprising an organizational area calculation procedure;
In the range selection procedure, in the pathological tissue image, the pixel having the maximum density weight value is selected, and when there are a plurality of the pixels, the distance weight value is the minimum among the pixels. If there are a plurality of such pixels, the pixel having the maximum weight value of the tissue region is selected from the pixels, and the specified range with the selected pixel as the upper left origin is The pathological tissue image photographing program according to claim 13, which is selected as an enlarged photographing range.   In the tissue region calculation procedure, a pixel having a maximum value of 64 or more among the luminances of R, G, and B in the pathological tissue image is selected, and the smaller one of the luminances of R and B is selected from the selected pixels. A pixel in which the value obtained by subtracting the luminance of G from the value of G is 10 or more is selected, and 255 is selected as the value obtained by taking the difference between the maximum value and the minimum value among the luminances of R, G, and B from the selected pixels. 15. The pathological tissue imaging program according to claim 14, wherein a pixel whose value obtained by dividing the multiplied value by the maximum value is 32 or more is selected, and the selected pixel is detected as a pixel located in the tissue region.

Images