Tissue image analysis device, tissue image analysis method, and program

Abstract

To provide a structure image analysis device, a method, and a program, which analyze a feature related to the form of fibers included in a structure image that does not have height information.SOLUTION: In a structure image analysis device, a processor 2 includes: an edge processing unit 201 configured to detect edges in a structure image including fibers and perform line segment approximation on the detected edges; a line segment pair identification unit 202 configured to draw a plurality of auxiliary lines in a line segment image showing approximated line segments, and identify a pair of line segments to be analyzed based on a change in luminance for a line segment intersecting an auxiliary line in a direction perpendicular to the line segment; and a feature amount calculation unit 203 configured to calculate a feature amount (e.g. the diameter) related to the fibers based on geometrical information of the identified pair of line segments.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a tissue image analysis device, a tissue image analysis method, and a program for analyzing a tissue image. [Background technology]

[0002] Patent Document 1 discloses that, in analyzing the morphology of fibers captured in a tissue image, which is an image of tissue, the diameter of the fibers is calculated by calculating the difference between the height of each position in the planar image of the fibers and the average height of the surface of the substrate. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2020-139239 Summary of the Invention [Problem to be solved by the invention]

[0004] However, Patent Document 1 states that height information for each position is required to calculate the fiber diameter, and that an atomic force microscope or other measuring means capable of obtaining height information is used. Therefore, no consideration is given to measuring means that cannot obtain height information, such as scanning electron microscopes and optical microscopes, and the applicable structure images are limited to those that include height information. Furthermore, since it is assumed that fibers do not overlap, no consideration is given to calculating the fiber diameter when overlapping occurs, and the applicable fibers are limited.

[0005] In view of the above circumstances, there is a demand for a technique that can analyze characteristics related to the morphology of fibers contained in tissue images that do not have height information. [Means for solving the problem]

[0006] One representative embodiment of the present invention is a tissue image analysis device comprising: an edge processing unit that detects edges in a tissue image containing fibers and performs line segment approximation on the detected edges; a line segment pair identification unit that draws a plurality of auxiliary lines in a line segment image representing the approximated line segments and identifies pairs of line segments to be analyzed based on luminance changes in a direction perpendicular to the line segments that intersect with the auxiliary lines; and a feature calculation unit that calculates feature amounts related to the fibers based on geometric information of the identified pairs of line segments.

[0007] One representative embodiment of the present invention is a tissue image analysis method that detects edges in a tissue image that includes fibers, performs line segment approximation of the detected edges, draws a plurality of auxiliary lines in a line segment image that represents the approximated line segments, identifies pairs of line segments to be analyzed based on luminance changes in a direction perpendicular to the line segments that intersect with the auxiliary lines, and calculates feature quantities related to the fibers based on geometric information of the identified pairs of line segments.

[0008] One representative embodiment of the present invention is a program for causing a computer to function as an edge processing unit that detects edges in a tissue image containing fibers and performs line segment approximation of the detected edges; a line segment pair identification unit that draws a plurality of auxiliary lines in a line segment image representing the approximated line segments and identifies pairs of line segments to be analyzed based on luminance changes in a direction perpendicular to the line segments that intersect with the auxiliary lines; and a feature calculation unit that calculates feature amounts related to the fibers based on geometric information of the identified line segment pairs. [Effects of the Invention]

[0009] According to a representative embodiment of the present invention, it is possible to analyze features related to the morphology of fibers contained in a tissue image that does not have height information. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a diagram showing a hardware configuration of a tissue image analyzing apparatus according to a first embodiment. [Figure 2] FIG. 2 is a functional block diagram of the tissue image analysis device. [Figure 3] FIG. 2 is a diagram showing an example of a processing flow in a tissue image analyzing apparatus. [Figure 4] FIG. 10 is a diagram for explaining an example of processing performed on a tissue image. [Figure 5] 10A and 10B are diagrams illustrating an example of generating diameter-corresponding line segments from paired detected lines; [Figure 6] FIG. 10 is a diagram for explaining the maximum and minimum diameters of fibers to be measured. [Figure 7] 10A and 10B are diagrams illustrating an example of a high line density and an example of a low line density. [Figure 8] FIG. 10 is a diagram showing the relationship between the degree of bending of a fiber and the length and number of approximate line segments. [Figure 9] FIG. 10 is a diagram illustrating an example of an analysis result. [Figure 10] FIG. 2 is a diagram showing an example of an input / output screen in the tissue image analyzing device. [Figure 11] FIG. 10 is a diagram for explaining analysis of a gap according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present invention will be described. Note that the present invention should not be construed as being limited to the description of the embodiments shown below. Those skilled in the art will easily understand that the specific configuration can be changed within the scope of the idea or purpose of the present invention.

[0012] In the configurations of the embodiments described below, the same reference numerals are used in common between different drawings for the same parts or parts having similar functions, and duplicated explanations may be omitted. When there are multiple elements having the same or similar functions, they may be described using the same reference numerals with different subscripts. However, when it is not necessary to distinguish between multiple elements, the subscripts may be omitted in the description.

[0013] The designations "first," "second," "third," etc. in this specification are used to identify components and do not necessarily limit the number, order, or content thereof. Furthermore, numbers used to identify components are used in different contexts, and numbers used in one context do not necessarily indicate the same configuration in another context. Furthermore, this does not prevent a component identified by a certain number from also serving the function of a component identified by another number.

[0014] In order to facilitate understanding of the invention, the position, size, shape, range, etc. of each component shown in the drawings etc. may not represent the actual position, size, shape, range, etc. Therefore, the present invention is not necessarily limited to the position, size, shape, range, etc. disclosed in the drawings etc.

[0015] One feature of the embodiment described in detail below is that edges in a tissue image are approximated by lines, and based on the change in brightness around the intersection of an auxiliary line drawn on the image and the approximated line, information on the direction of fiber existence is added to the line, thereby detecting fibers and determining the diameter of the fibers, etc.

[0016] (Embodiment 1) A description will be given of a tissue image analyzing device according to a first embodiment of the present invention. The tissue image analyzing device includes an edge processing unit that detects edges in a tissue image including fibers and performs line segment approximation on the detected edges, a line segment pair specifying unit that draws a plurality of auxiliary lines in a line segment image representing the approximated line segments and specifies pairs of line segments to be analyzed based on luminance changes in a direction perpendicular to the line segments that intersect with the auxiliary lines, and a feature amount calculation unit that calculates feature amounts related to the fibers based on geometric information of the specified line segment pairs.

[0017] Figure 1 is a diagram showing the hardware configuration of a tissue image analyzing device according to embodiment 1. As shown in Figure 1, tissue image analyzing device 1 is, for example, a computer, and is configured by a processor 2, memory 3, storage device 4, and network adapter 5, all of which are connected via a system bus 6 so as to be capable of transmitting and receiving signals. Tissue image analyzing device 1 is also connected via a network 9 to a tissue image capturing device 10 or a tissue image database 11 so as to be capable of transmitting and receiving signals, and is also connected to a display device 7 and an input device 8. Here, "capable of transmitting and receiving signals" refers to a state in which signals can be transmitted and received between them or from one device to the other, whether electrically or optically, wired or wirelessly.

[0018] The processor 2 is a device that controls the operation of each component and executes programs stored in the storage device 4. The memory 3 stores the programs executed by the processor 2 and the progress of arithmetic processing. The storage device 4 is a device that stores the programs executed by the processor 2 and data necessary for executing the programs, and is specifically, for example, an HDD (Hard Disk Drive) or an SSD (Solid State Drive). The network adapter 5 is used to connect the tissue image analysis device 1 to a network 9 such as a LAN (Local Area Network), a telephone line, or the Internet. Various data handled by the processor 2 may be transmitted and received from outside the tissue image analysis device 1 via the network 9 such as a LAN.

[0019] The display device 7 is a device that displays the analysis results of the tissue image analyzer 1, and specifically includes, for example, a liquid crystal display or a touch panel. The input device 8 is an operation device that allows an operator to issue operation instructions to the tissue image analyzer 1, and specifically includes, for example, a keyboard, a mouse, a touch panel, etc. A pointing device such as a trackpad or a trackball may be used instead of a mouse.

[0020] The tissue image capturing device 10 is a device for capturing tissue images, such as a scanning electron microscope or an optical microscope. The tissue image database 11 is a database system that stores tissue images captured by the tissue image capturing device 10.

[0021] Next, a functional block diagram of the tissue image analyzing device 1 according to the first embodiment will be described.

[0022] Fig. 2 is a functional block diagram of a tissue image analyzer. Each functional block shown in Fig. 2 may be realized by dedicated hardware using an ASIC (Application Specific Integrated Circuit) or an FPGA (Field-Programmable Gate Array), or may be realized by a computer, i.e., a processor 2, executing a program. In the following explanation, the tissue image analyzer 1 is a computer, and the computer functions as each functional block by executing a program.

[0023] The tissue image analysis device 1 includes, as functional blocks, an edge processing unit 201, a line segment pair identification unit 202, and a feature amount calculation unit 203. The storage device 4 stores tissue images captured by the tissue image capturing device 10. Each functional block will be described below.

[0024] The edge processing unit 201 performs various image processing on the edges of the tissue image. This image processing includes edge detection processing to detect edges and line segment approximation processing to approximate edges with line segments. Hereinafter, these approximated line segments will be referred to as approximate line segments or simply line segments.

[0025] The line segment pair identifying unit 202 performs processing to identify line segment pairs corresponding to locations suitable for measuring fiber diameters. This processing includes a direction analysis processing in which an auxiliary line is drawn on the line segment image representing the approximate line segment, the position of the intersection between the auxiliary line and the line segment is identified, and information indicating the direction of fiber existence is added to the line segment based on the brightness change around the position corresponding to the identified position in the tissue image, and a line segment pair detection processing in which the line segment pair to be analyzed is detected based on the information added to the line segment.

[0026] The feature calculation unit 203 calculates feature quantities such as the diameter and angle of the fiber based on the geometric information of the detected line segment pairs, and outputs the analysis results. The analysis results may be, for example, various feature quantities as well as a histogram of the feature quantities.

[0027] An example of the processing flow in the tissue image analyzer 1 will now be described step by step with reference to the drawings.

[0028] Fig. 3 is a diagram showing an example of the processing flow in the tissue image analysis apparatus 1. As shown in Fig. 3, in step S301, the edge processing unit 201 acquires a tissue image by reading it from the storage device 4 or receiving it from outside the tissue image analysis apparatus 1 via the network adapter 5. The tissue image is assumed to be an image without height information, such as a grayscale image acquired by an electron microscope, a color image acquired by an optical microscope, or a multi-channel image acquired by a multispectral camera.

[0029] In step S302, the edge processing unit 201 detects edges from the tissue image and creates a binarized image based on the detection result.

[0030] Fig. 4 is a diagram for explaining an example of processing performed on a tissue image. For example, when the tissue image is an input image 401 shown in the first row of the left column of Fig. 4, an edge-extracted image 402 shown in the second row of the left column of Fig. 4 is created as a binarized image, and detected edges 4001 are displayed as white lines.

[0031] Any method can be used to detect edges. For example, edges can be detected by differentiating spatial brightness changes in the tissue image and binarizing the obtained values, or by using a method known as the Canny algorithm. Brightness changes can generally be thought of as changes in pixel values. Here, the range, i.e., kernel size, used to differentiate spatial brightness changes can be specified to adjust the spread of edges to be detected.

[0032] The generated binary image is a monochrome image in which edge portions corresponding to the outlines of the fibers correspond to white and other portions correspond to black. Note that a monochrome image in which edge portions corresponding to the outlines of the fibers correspond to black and other portions correspond to white may also be generated as the binary image.

[0033] In step S303, the edge processing unit 201 performs line segment approximation of the edges of the binarized image in which the edges have been detected, and associates the obtained approximate line segment with the tissue image. For example, in the line segment approximation image 403 shown in the third row of the left column in FIG. 4, an approximate line segment 4002 corresponding to the edge, as exemplified by a black dashed line, is obtained. Any method can be used to approximate the edge using a line segment, and for example, approximation may be performed using a method known as a Hough transform or a method known as a fast line detector.

[0034] In step S304, the line segment pair identification unit 202 draws a plurality of auxiliary lines in the line segment image representing the approximate line segments corresponding to the edges. The plurality of auxiliary lines can be obtained, for example, by drawing auxiliary lines while changing at least one of the distance interval and the angle interval by a fixed amount.

[0035] In step S305, the line segment pair identification unit 202 adds information indicating the fiber presence direction to the approximate line segment that intersects with the auxiliary line. The fiber presence direction relative to the approximate line segment can be determined by comparing the brightness or color on both sides of the approximate line segment in the tissue image, i.e., on one side and the other side of the approximate line segment in a direction perpendicular to the longitudinal direction of the approximate line segment. An enlarged image 404, an enlarged partial region R404, which is a partial region of the line segment approximate image 403, is shown in the first row of the right column of Figure 4. In the enlarged image 404, for the approximate line segment that intersects with the dotted auxiliary line 4003, the brightness on the tissue image on both sides of the approximate line segment is compared, and the brighter side, i.e., the side with higher brightness, can be determined as the fiber presence direction. In the enlarged image 404 in Figure 4, the fiber presence direction is represented by a white arrow.

[0036] For ease of understanding, only one auxiliary line 4003 is shown in the enlarged image 404 of Figure 4, but in reality, multiple auxiliary lines at different positions or angles are drawn on the line segment image, and information indicating the fiber direction is added to the multiple approximate line segments. Furthermore, if the tissue image is a color image, the fiber direction can be determined by similar processing using only a specific color channel or by converting to grayscale.

[0037] In step S306, the line segment pair identifying unit 202 sets a pair of line segments sandwiching a fiber (hereinafter also referred to as a line segment pair) from among the multiple approximate line segments as an analysis point suitable for analysis. To identify the line segment pair, a method is used in which the fiber directions at the intersection points of the auxiliary lines and the approximate line segments are compared, and a combination of line segments whose fiber directions are opposite to each other is identified as a line segment pair.

[0038] For example, in the case of the enlarged image 404 shown in the first row of the right column in Fig. 4, when examining the auxiliary line 4003 shown by the black dotted line from the left side, the fiber directions at the three points where the auxiliary line 4003 intersects with the approximation line segments are rightward, rightward, and leftward, respectively, relative to the drawing. Of the line segments corresponding to these three intersections, the combination of line segments whose fiber directions are opposite, i.e., the line segment corresponding to the second intersection from the left and the line segment corresponding to the third intersection from the left, can be identified as a line segment pair suitable for analysis.

[0039] As mentioned above, multiple auxiliary lines can be drawn on a line segment image, and the distance intervals and rotation angle intervals of the auxiliary lines can be set. When drawing auxiliary lines by varying the distance intervals and angle intervals by a fixed amount, the smaller the amount of change, the easier it is to identify more pairs on the image, but this also increases processing time, so the amount of change may be adjusted appropriately. For example, the distance interval of the auxiliary lines is preferably set to 0.1 to 50% of the length of the longest side of the image. This corresponds to 1 to 500 pixels when the longest side of the image is 1,000 pixels long. The rotation angle interval of the auxiliary lines is preferably set to between 1 and 90 degrees.

[0040] An example of a case where line segment pairs are identified using multiple auxiliary lines is the line segment pair detection image 405 shown in the second row on the right side of Fig. 4. In this line segment pair detection image 405, lines indicating locations identified as line segment pairs, i.e., pair detection lines 4004 which are lines connecting the line segments of the identified line segment pairs, are represented by black straight lines.

[0041] In step S307, the feature calculation unit 203 calculates the feature of the tissue based on the processing results up to this point.

[0042] First, in calculating the fiber diameter value, the angles of the two approximate line segments located at both ends of the pair detection line drawn at the analysis location identified as a line segment pair are calculated as angles with respect to the pair detection line, as exemplified in the line segment pair detection image 405. Next, based on these calculated angles, the pair detection line is converted into a line segment corresponding to the diameter, thereby generating a diameter corresponding line segment.

[0043] FIG. 5 is a diagram illustrating an example of generating a radially corresponding line segment from a paired detection line. As shown in FIG. 5, for example, the average of the angles of the alternate angles (the combination of upper right corner A1 and lower left corner A2 or the combination of upper left corner B1 and lower right corner B2) between two approximation line segments 502 located at both ends of the paired detection line 501 is calculated. The paired detection line 501 is then rotated by an angle obtained by subtracting the average angle from 90 degrees. In this case, the counterclockwise direction is positive for the combination of angles A1 and A2, and the clockwise direction is positive for the combination of angles B1 and B2. Furthermore, both ends of each approximation line segment 502 are extended, and the lengths of both ends of the rotated paired detection line 501 are adjusted so that they are tangent to the extended approximation line segments. Through these processes, the paired detection line 501 is converted into a radially corresponding line segment, and the converted line segment is designated as a radially corresponding line segment 503.

[0044] The length of the diameter corresponding line segment can then be determined as the diameter of the fiber. In the diameter measurement image 406 shown in the third row on the right side of Figure 4, the diameter corresponding line segment 4005 is represented by a black dashed line.

[0045] Furthermore, the input tissue image may contain noise or other contaminants. If noise or other contaminants are mixed into the tissue image, short paired detection lines may be generated due to the noise or other contaminants, which may adversely affect the measurement of the fiber diameter. To cope with such a situation, an allowable range may be set for the diameter of the fiber to be measured, and if the measured diameter value falls outside the allowable range, the diameter may be excluded from the analysis results, or the line segment pair used to calculate the diameter may be excluded from the analysis targets.

[0046] Fig. 6 is a diagram for explaining the maximum and minimum values ​​of the diameter of the fiber to be measured. As shown in Fig. 6, by setting a minimum value 601 and a maximum value 602 of the diameter of the fiber to be measured, it is possible to narrow down the length of the diameter 603 within the range to be measured.

[0047] For example, by appropriately setting the minimum allowable diameter value, it is possible to eliminate short pair detection lines caused by minute noise in the image. Furthermore, by appropriately setting the maximum allowable diameter value, it is possible to eliminate long pair detection lines caused by incorrect pairs. The minimum diameter value is preferably in the range of 0.1 to 50% of the longest side of the image, and the maximum diameter value is preferably in the range of 1 to 100% of the longest side of the image. In this case, the maximum diameter value should be greater than the minimum diameter value. That is, if the length of the longest side of the image is 1000 pixels, it is preferable that the minimum diameter be in the range of 1 to 500 pixels and the maximum diameter be in the range of 10 to 1000 pixels.

[0048] Furthermore, the fiber angle can be calculated based on the longitudinal angles of the line segments that make up the identified line segment pairs. Then, by calculating the fiber angles for all line segment pairs, the distribution of fiber angles can be calculated.

[0049] Furthermore, the ratio of the length of the line segments actually identified as line segment pairs on the auxiliary lines to the length of the auxiliary lines used to identify the line segment pairs, i.e., the line density, increases as the fibers become denser in the image. Therefore, a value corresponding to the degree of fiber density can be calculated from the line density.

[0050] FIG. 7 is a diagram showing an example of a high line density and an example of a low line density. FIG. 7 illustrates a schematic diagram 701 showing a high line density and a schematic diagram 702 showing a low line density. As shown in the schematic diagrams 701 and 702, when auxiliary lines shown as white dotted lines are drawn in an image, the proportion of the length occupied by the portion identified as the pair detection line shown as a solid white line increases as the fibers shown in dark colors become denser. In other words, this proportion is higher in the case of the schematic diagram 701, in which the fibers are denser. This makes it possible to calculate a value corresponding to the degree of fiber density.

[0051] In addition to the above, a value corresponding to the degree of bending of the fiber can also be calculated from the distribution of the lengths of the approximated line segments exemplified in the line segment approximation image 403 in FIG.

[0052] FIG. 8 is a diagram showing the relationship between the degree of fiber curvature and the length and number of approximation line segments. FIG. 8 illustrates a schematic diagram 801 of approximation line segments for an edge with a large curvature and a schematic diagram 802 of approximation line segments for an edge with a small curvature. In FIG. 8, edges are represented by thick dark lines, and approximation line segments are represented by straight dotted lines with black circles at their endpoints. As shown in schematic diagrams 801 and 802, the greater the edge curvature, the greater the number of approximation line segments and the shorter their lengths become. In other words, it is possible to grasp the degree of fiber curvature based on information such as whether there are more short or long approximation line segments in the distribution of their lengths.

[0053] In step S308, the feature calculation unit 203 outputs the analysis results, i.e., the calculated results, of each feature. For example, the feature calculation unit 203 may output each feature calculated in step S307 as is, or may output information representing a histogram of the feature.

[0054] Fig. 9 is a diagram showing an example of the analysis results, which, from top to bottom, show a histogram 901 of diameters, a histogram 902 of angles, a histogram 903 of lengths of approximate line segments that form line segment pairs, and a histogram 904 of line density.

[0055] FIG. 10 is a diagram showing an example of an input / output screen in the tissue image analyzer 1. In the input / output screen 1001 shown in FIG. 10, the file name of an input image showing a fibrous object is specified in the input image specification field. Note that multiple images stored in a folder may be specified by clicking a button labeled "Select Folder." In the parameter setting field, various parameters for performing analysis processing are set. The various parameters include, for example, the kernel size for edge extraction, the distance interval of auxiliary lines, the angle interval for rotation of auxiliary lines, the minimum diameter to be measured, and the maximum diameter to be measured.

[0056] After specifying the input image and setting the parameters, clicking the "Run Analysis" button executes the analysis, and the analysis results are displayed in the preview area. Examples of displayed analysis results include an edge detection image, a line segment approximation image, a line segment pair detection image, and a diameter measurement image, as shown in Figure 10. Examples of displayed analysis results include a diameter histogram (analysis result of fiber diameter), a histogram of approximate line segment angles (analysis result of fiber angle), a line segment pair length histogram (analysis result of the length of the line segments that make up the line segment pair, i.e., the line segment pair length), and a line density histogram (analysis result of fiber density), as shown in Figure 10. These analysis results can be saved by clicking "Save Analysis Results." If "Yes" is selected for the setting to output detailed numerical values ​​from the analysis results, a detailed numerical value file of the analysis results will also be saved.

[0057] As described above, according to the first embodiment, even when fibers appearing in a tissue image overlap, analysis results regarding diameter and angle are output without any problems, making it possible to analyze the morphological characteristics of fibers contained in a tissue image that does not contain height information. Furthermore, the analysis results, i.e., the feature quantities, may be incorporated into machine learning or the like and used to optimize manufacturing methods and performance. Furthermore, because the analysis of tissue images is performed by computation, it is possible to achieve faster and more labor-saving results than when equivalent tasks are performed by humans.

[0058] (Embodiment 2) In the first embodiment, the case where the analysis of fibers is the main focus has been described. On the other hand, there are cases where the characteristics of regions where fibers are not present, i.e., void portions, are important. Therefore, in the second embodiment, the analysis of voids will also be described. Note that the difference from the first embodiment is whether or not voids are analyzed, so the analysis of voids will be mainly described. The hardware configuration and functional block diagram of the tissue image analyzer 1 are the same as those in the first embodiment, so their description will be omitted, and the processing flow is similar to that in the first embodiment, so the description will be simplified.

[0059] FIG. 11 is a diagram illustrating analysis of voids according to the second embodiment. To analyze voids, it is first necessary to identify the voids. However, the contrast and brightness of an input image are not necessarily constant, and the proportion of voids in the image is not necessarily constant. Therefore, binarization methods that determine a threshold value based on the distribution of brightness across the entire image, such as the method known as Otsu's binarization, may not be able to properly extract voids.

[0060] Therefore, after obtaining a line corresponding to the fiber diameter, i.e., a pair detection line, using the method of embodiment 1, the position of the pair detection line and the brightness on the input image corresponding to that pair detection line are identified. Then, based on the identified brightness, areas of pixels in the input image, i.e., the tissue image, whose brightness differs from the identified brightness by a certain level are painted white, thereby identifying areas where fibers exist. Conversely, areas where fibers do not exist, i.e., areas different from the white-painted areas, can be identified as void areas. In the void image 1101 shown in the upper part of Figure 11, void areas are represented in black, and areas where objects exist in white.

[0061] Next, the voids can be analyzed using a technique known as the intercept method or intercept method. In the middle of Figure 11, a void continuation 1103, shown as a solid line, is shown on a void auxiliary line 1102, shown as a dotted line. The length of the void continuation on the void auxiliary line 1102 is calculated as a feature value related to the fiber. That is, the length of the void continuation is obtained as an index related to the size of the void in the image. By drawing void auxiliary lines at various angles and tallying the lengths of the void continuations, the size of the voids in the image can be determined from the distribution of these lengths.

[0062] 11 illustrates an example of a gap length histogram 1104 as an analysis result of the gap portion. The feature amount calculation unit 203 may be configured to output information representing such a gap portion length histogram.

[0063] As described above, according to the second embodiment, even if the contrast, brightness, or fiber ratio of the input image is not constant, it is possible to analyze the voids contained in the tissue image and output the analysis results of the void lengths. This allows obtaining feature quantities not only for the fibers but also for the voids.

[0064] The tissue image analysis device and tissue image analysis method according to the embodiment of the present invention have been described above. The present invention is not limited to the above-described embodiment, and the components can be modified and embodied without departing from the spirit and scope of the invention.

[0065] For example, a program for causing a computer to function as each functional block of the tissue image analyzer is also an embodiment of the present invention. Furthermore, a tangible computer-readable storage medium on which the program is non-temporarily stored is also an embodiment of the present invention. [Explanation of symbols]

[0066] 1: Tissue image analysis device, 2: Processor, 3: Memory, 4: Storage device, 5: Network adapter, 6: System bus, 7: Display device, 8: Input device, 9: Network, 10: Tissue image capture device, 11: Tissue image database, 201: Edge processing unit, 202: Line segment pair identification unit, 203: Feature amount calculation unit, 401: Input image, 402: Edge extracted image, 403: Line segment approximation image, 404: Enlarged image, 405: Line segment pair detection image, 406: Diameter measurement image, 4001: Edge, 4002: Approximation line segment, 4003: Auxiliary line, 4004: Pair detection line, 4005: Diameter corresponding line segment, 501: Pair detection line, 502: Approximation line segment, 503: diameter corresponding line segment, 601: minimum diameter to be measured, 602: maximum diameter to be measured, 603: diameter in the range to be measured, 701: schematic diagram when line density is high, 702: schematic diagram when line density is low, 801: schematic diagram of approximation line segment for edge with large curvature, 802: schematic diagram of approximation line segment for edge with small curvature, 901: diameter histogram, 902: angle histogram, 903: histogram of length of paired approximation line segment, 904: line density histogram, 1001: input / output screen, 1101: gap image, 1102: gap auxiliary line, 1103: gap continuation portion, 1104: gap length histogram

Claims

1. an edge processing unit that detects edges in a tissue image including fibers and performs line segment approximation on the detected edges; a line segment pair specifying unit that draws a plurality of auxiliary lines in a line segment image that represents the approximated line segments, and specifies pairs of the line segments to be analyzed based on a luminance change in a direction perpendicular to the line segments that intersect with the auxiliary lines; a feature amount calculation unit that calculates a feature amount related to the fiber based on geometric information of the identified pair of line segments, Tissue image analysis device.

2. 2. The tissue image analysis device according to claim 1, the line segment pair identifying unit draws a plurality of the auxiliary lines in the tissue image by changing at least one of a distance interval and an angle interval, adds information on the fiber existence direction to the line segments intersecting with the auxiliary lines based on a brightness change in a direction perpendicular to a longitudinal direction of the line segments, and identifies the line segments whose added existence directions are opposite to each other as a pair of the line segments. Tissue image analysis device.

3. 2. The tissue image analysis device according to claim 1, the feature calculation unit calculates an angle of each of the two line segments with respect to a pair detection line that is a line connecting the two line segments constituting the identified pair, converts the pair detection line into a diameter corresponding line segment that corresponds to the diameter of the fiber based on the calculated angle, and calculates the diameter of the fiber as the feature based on the length of the diameter corresponding line segment. Tissue image analysis device.

4. 2. The tissue image analysis device according to claim 1, the feature calculation unit calculates an angle of the fiber as the feature based on an angle in a longitudinal direction of the line segments constituting the pair; Tissue image analysis device.

5. 2. The tissue image analysis device according to claim 1, the feature amount calculation unit calculates, as the feature amount, a value corresponding to a density of the fibers based on a ratio of a length between the line segments identified as the pair on the auxiliary line to a length of the auxiliary line. Tissue image analysis device.

6. 2. The tissue image analysis device according to claim 1, the feature amount calculation unit calculates, as the feature amount, a value corresponding to a degree of bending of the fiber based on the distribution of the approximated lengths of the line segments. Tissue image analysis device.

7. 2. The tissue image analysis device according to claim 1, the feature amount calculation unit specifies a luminance on the tissue image corresponding to a pair detection line that is a line connecting two line segments that make up the specified pair, specifies an area in the tissue image that is different from an area of ​​luminance whose difference from the specified luminance is within a certain level, as a void portion, and calculates a length of the void portion on the void auxiliary line as the feature amount. Tissue image analysis device.

8. 2. The tissue image analysis device according to claim 1, the feature calculation unit outputs information representing a histogram of the feature. Tissue image analysis device.

9. 4. The tissue image analyzer according to claim 3, the line segment pair specifying unit excludes the calculated diameter of the fiber or the pair used to calculate the diameter from an analysis target when the calculated diameter of the fiber is outside a set allowable range; Tissue image analysis device.

10. 10. The tissue image analysis device according to claim 1, The tissue image is an image obtained by an electron microscope, an optical microscope, or a multispectral camera. Tissue image analysis device.

11. A tissue image analysis method using a tissue image analyzer, comprising: Detecting edges in a tissue image including fibers, and performing line segment approximation on the detected edges; drawing a plurality of auxiliary lines in the line segment image representing the approximated line segments, and identifying pairs of the line segments to be analyzed based on a change in luminance of the line segments that intersect with the auxiliary lines in a direction perpendicular to the line segments; calculating a feature amount of the fiber based on geometric information of the identified pair of line segments; Tissue image analysis method.

12. Computer, an edge processing unit that detects edges in a tissue image including fibers and performs line segment approximation on the detected edges; a line segment pair specifying unit that draws a plurality of auxiliary lines in a line segment image that represents the approximated line segments, and specifies pairs of the line segments to be analyzed based on a luminance change in a direction perpendicular to the line segments that intersect with the auxiliary lines; and a program for causing the program to function as a feature calculation unit that calculates feature amounts related to the fiber based on geometric information of the identified pair of line segments.

Images