Image evaluation method, image evaluation device, computer program, and non-transitory computer-readable medium

Abstract

Image data corresponding to an ultrasonic image in which a pelvic region of a subject is reflected is acquired. A plurality of feature points related to a given portion are extracted by applying image processing to the image data. An index value associated with a lower urinary tract function and / or a pelvic floor function is calculated on the basis of the plurality of feature points.

Description

Technical Field

[0001] This disclosure relates to a method and apparatus for evaluating indicators associated with at least one of lower urinary tract function and pelvic floor function based on ultrasound images of the pelvic region of a subject, a computer program executable by a processor mounted on the apparatus, and a non-transitory computer-readable medium storing the computer program. Background Technology

[0002] Patent Document 1 discloses a device capable of acquiring ultrasound images of the pelvic region of a subject. These ultrasound images can be used, for example, to evaluate lower urinary tract function after total prostatectomy, and for pelvic floor muscle training to suppress incontinence caused by decreased pelvic floor function after childbirth, etc. However, the evaluation of these ultrasound images depends on the observer's subjectivity and experience.

[0003] Prior art literature

[0004] Patent documents

[0005] Patent Document 1: JP Patent Application Publication No. 2018-504185 Summary of the Invention

[0006] -The problem the invention aims to solve-

[0007] The aim is to provide a method that can objectively evaluate indicators associated with at least one of lower urinary tract function and pelvic floor function based on ultrasound images of the subject's pelvic region.

[0008] -Methods used to solve problems-

[0009] One example of the method provided by this disclosure is an image evaluation method executed by at least one computing device, which acquires image data corresponding to an ultrasound image of the pelvic region of the subject, extracts multiple feature points related to a given location by applying image processing to the image data, and calculates an index value that is associated with at least one of lower urinary tract function and pelvic floor function based on the multiple feature points.

[0010] One example of the method provided by this disclosure is an image evaluation device comprising: an interface that receives image data corresponding to an ultrasound image of the pelvic region of a subject; and a processor that extracts multiple feature points related to a given location by applying image processing to the image data, and calculates an index value based on the multiple feature points that is associated with at least one of lower urinary tract function and pelvic floor function.

[0011] One example of the method provided by this disclosure is a computer program that can be executed by a processor mounted on an image evaluation device. By executing the computer program, the image evaluation device performs the following processing: receiving image data corresponding to an ultrasound image of the pelvic region of the subject; extracting multiple feature points related to a given location by applying image processing to the image data; and calculating an index value that is associated with at least one of lower urinary tract function and pelvic floor function based on the multiple feature points.

[0012] A non-transitory computer-readable medium storing the aforementioned computer program is also one of the embodiments provided by this disclosure.

[0013] Based on the structures described in the above examples, by simply inputting an ultrasound image of the subject's pelvic region, the processing steps up to calculating an index value that correlates the subject with at least one of lower urinary tract function and pelvic floor function can be automated. Since this index value is not dependent on the subjective, empirical observation of the person viewing the ultrasound image, an objective evaluation environment can be provided. Attached Figure Description

[0014] Figure 1 The diagram illustrates the organs contained in the male pelvic region.

[0015] Figure 2 This example illustrates the state of the patient after undergoing a total prostatectomy.

[0016] Figure 3 An example of an image evaluation method is provided in one embodiment.

[0017] Figure 4 Example Figure 3 The detailed process of image processing in [the context of the image processing].

[0018] Figure 5 An example of an ultrasound image projected onto the pelvic region of a subject.

[0019] Figure 6 Example in Figure 5 The state of the region of interest was determined within the ultrasound image.

[0020] Figure 7 Example of extracting only Figure 6 The image of the region of interest.

[0021] Figure 8 Example of Figure 7 The image is a binary image.

[0022] Figure 9 Example of Figure 8 The image is processed using edge extraction.

[0023] Figure 10 Example of Figure 9 The image is an image that has undergone the Hough transform.

[0024] Figure 11 This is a diagram used to calculate the angle of the bladder neck.

[0025] Figure 12 This is a diagram used to calculate the angle of the bladder neck.

[0026] Figure 13 This is a diagram used to calculate the angle of the bladder neck.

[0027] Figure 14 The example shows the angle of the bladder neck.

[0028] Figure 15 This is a diagram used to illustrate the angle of the bulbous urethra.

[0029] Figure 16 An example is an ultrasound image of the subject's urethra.

[0030] Figure 17 This is a diagram used to illustrate the process of calculating the angle of the bulbous urethra.

[0031] Figure 18 This is a diagram used to illustrate the process of calculating the angle of the bulbous urethra.

[0032] Figure 19 This is a diagram used to illustrate the process of calculating the angle of the bulbous urethra.

[0033] Figure 20 It is a diagram used to illustrate the distance between the lower edge of the pubic bone and the anorectal angle.

[0034] Figure 21 This diagram illustrates the process of calculating the distance between the lower edge of the pubic bone and the anorectal angle.

[0035] Figure 22 The functional structure of the image evaluation device involved in the first embodiment is illustrated. Detailed Implementation

[0036] Referring to the accompanying drawings, and described in detail below, an embodiment of a method for evaluating indicators associated with at least one of lower urinary tract function and pelvic floor function based on ultrasound images of the subject's pelvic region is described.

[0037] Figure 1 The organs contained in the male pelvic region are illustrated schematically. Figure 2 Examples illustrate the state after undergoing total prostatectomy, a known radical treatment for prostate cancer.

[0038] In this surgery, the prostate gland (1) is removed, and the bladder (2) and urethra (3) are connected. As a result of this surgery, the bladder (2) is pulled downwards, sometimes creating a funnel-shaped enlargement of the bladder neck (4). Because urine in the bladder (2) can more easily flow into the urethra (3) due to this change in shape, it has been indicated that the formation of the funnel can be a mechanism for urinary incontinence. Urinary incontinence is a condition associated with decreased function of the lower urinary tract.

[0039] The characteristic feature of funnel formation is that the angle θ of the bladder neck becomes smaller than before surgery (180°). In cases where a funnel is considered to have formed (where the funnel angle is non-zero), the angle θ of the bladder neck corresponds to the degree of funneling. That is, the larger the angle θ, the greater the degree of funneling, and the higher the presumed severity of lower urinary tract dysfunction. Lower urinary tract function is evaluated by visually confirming ultrasound images of the pelvic region of the subject by a medical professional.

[0040] refer to Figure 3 This paper describes an image evaluation method according to one embodiment. The method is based on the idea that the angle θ of the bladder neck can serve as an indicator for quantitatively evaluating the function of the lower urinary tract. In this method, the aforementioned indicator is automatically calculated based on ultrasound images of the subject's pelvic region.

[0041] First, image data corresponding to the ultrasound image projected onto the subject's pelvic region is acquired (STEP1). The ultrasound image can be acquired, for example, by touching the ultrasound probe from the subject's groin.

[0042] Next, image processing (STEP2) is performed on the image data. Image processing is a preparatory process to enable the automatic calculation of the bladder neck angle θ. (Reference) Figure 4 This will illustrate the detailed process of image processing.

[0043] First, a region containing at least the bladder neck is extracted from the ultrasound image of the pelvic region as the region of interest (STEP21). This process is performed by applying a well-known object detection algorithm to the image data obtained in STEP1.

[0044] In this embodiment, YOLO (You Only Look Once) is used as the object detection algorithm. YOLO identifies regions with a probability higher than a given threshold that are mapped into the bladder neck as regions of interest. In creating the deep learning model that forms the basis of YOLO, teaching data annotated with the location of the bladder neck in ultrasound images mapped into the pelvic region is used.

[0045] Figure 5Example: Ultrasound image 11 of the subject's pelvic region. In the same image, one should note the vertical direction and... Figure 2 Conversely, the bladder neck is projected onto the upper part of the bladder. Figure 6 This example illustrates how the state of region of interest A was determined within ultrasound image 11 using YOLO. Figure 7 Example: Extracted image 12 is extracted from ultrasound image 11, focusing only on region A of interest.

[0046] Next, the extracted image 12 is binarized. Figure 4 (STEP 22). Each of the multiple pixels constituting the extracted image 12 has any one of a number of gray values ​​that can represent the shades of gray. In the binarization process, the gray value of each pixel is transformed into either a value corresponding to white or a value corresponding to black.

[0047] For example, if the grayscale value of a pixel is below a threshold, it is transformed into a value corresponding to black; if it is above the threshold, it is transformed into a value corresponding to white. The threshold can be a predetermined fixed value or a variable value determined by discriminant analysis. Figure 8 Example: Binarized image 13 is obtained by applying binarization processing to extracted image 12.

[0048] Next, edge extraction processing is performed on the binarized image 13. Figure 4 (STEP 23). Edge extraction is a process that emphasizes areas where the difference in grayscale values ​​between adjacent pixels is large. As can be seen from the binarized image 13, the grayscale values ​​change dramatically at the boundary between the bladder lumen and the inner wall. Therefore, edge extraction is used to emphasize the boundary portion that forms part of the bladder neck.

[0049] As methods for edge extraction, the Sobel method and the Laplacian method are known. In this example, the Laplacian method, which can extract edges more smoothly while suppressing the influence of noise, is used. Figure 9 An example is shown: edge-extracted image 14 is obtained by applying edge extraction processing to binarized image 13.

[0050] Next, the edge-extracted image 14 is processed by Hough transform. Figure 4 (STEP 24). The Hough transform is a process for detecting straight lines in an input image. Figure 10 Example: Line detection image 15 obtained by applying Hough transform to edge extraction image 14. It shows numerous lines arranged along the boundary between the bladder's lumen and its inner wall.

[0051] Additionally, for ease of understanding, at least binarized image 13, edge-extracted image 14, and line-detected image 15 are illustrated, but visualization is not necessarily required for the user. Each process can be described as a transformation process performed on the data corresponding to each image without accompanying visualization.

[0052] like Figure 3 As illustrated, the next step is to calculate the index value (STEP3). The angle θ of the subject's bladder neck is calculated based on the line detection image 15 described above, and this angle is used as the index value.

[0053] exist Figure 11 The diagram illustrates a graph obtained by using the midpoints of the various lines detected in line detection image 15 as representative points and plotting the positions of these representative points within the same image. The horizontal axis represents the horizontal coordinates in line detection image 15, and the vertical axis represents the vertical coordinates within the same image.

[0054] Next, clustering is applied to the many plotted points obtained as described above. Specifically, the slope values ​​of the lines that form the basis for each plotted point in the line detection image 15 are referenced. For example... Figure 12 As illustrated, points obtained from lines with positive slope values ​​are classified into cluster C1. Points obtained from lines with negative slope values ​​are classified into cluster C2. Figure 12 In the image 15, the vertical axis represents the horizontal coordinate of the line detection image, and the vertical axis represents the slope value.

[0055] In addition, points whose absolute slope value is above the threshold are classified as cluster C1, and points whose slope value is below the threshold are classified as cluster C12. Similarly, points whose absolute slope value is above the threshold are classified as cluster C2, and points whose slope value is below the threshold are classified as cluster C22.

[0056] Figure 13 Example from Figure 11 The illustrated plot points will be classified as excluding plot points in clusters C12 and C22. In other words, plot points with an absolute slope value above a threshold will be retained.

[0057] Line L1 is obtained by applying linear regression to multiple plotted points belonging to cluster C12, weighted according to the lengths of lines obtained through Hough transform. Line L1 is then compared with... Figure 5 The extracted image 12 corresponds to a straight line that approximates the inner wall of the bladder neck with a positive slope. Similarly, line L2 is obtained by applying linear regression to multiple plotted points belonging to cluster C22. Line L2 corresponds to the line that will be... Figure 5The extracted image 12 corresponds to a straight line that roughly corresponds to the inner wall of the bladder neck with a negative slope. The angle θ formed by the straight lines L1 and L2 is determined as the angle of the bladder neck.

[0058] The above clustering approach, which considers not only the positive or negative value of the slope but also the absolute value of the slope, is based on the insight that the bladder neck that forms the funnel has a shape that includes parts with relatively large slopes and parts with relatively small slopes.

[0059] Figure 11 The illustrated line L1' was obtained by applying linear regression to all plotted points belonging to cluster C1. Similarly, the line L2' was obtained by applying linear regression to all plotted points belonging to cluster C2. These lines cannot be said to adequately approximate the actual shape of the bladder neck, and the angle θ' between lines L1' and L2' is also difficult to say to correspond to the actual angle of the bladder neck.

[0060] pass Figure 11 and Figure 13 The comparison shows that by focusing on the relatively steep part of the bladder neck, a value closer to the actual angle of the bladder neck can be calculated.

[0061] The clustering described above can be achieved by applying at least one well-known method. Examples of well-known methods include k-means and Gaussian mixture distribution. Furthermore, it is not necessary to focus on both the slope and magnitude of the line for clustering. For example, one can focus only on the magnitude of the slope and perform at least one clustering operation.

[0062] like Figure 3 As illustrated, the next step is to process the output of the result data (STEP4). The result data is configured to include at least the data corresponding to the angle of the bladder neck calculated as described above.

[0063] Figure 14 This is an example of how the result data is output. In this example, the calculated angle of the bladder neck is displayed on the display device. That is, the result data is configured to cause the display device to display the value of that angle. In this example, image 16, representing the basis for calculating that angle, is also displayed on the display device. Image 16 involved in this example is related to... Figure 10 The image corresponding to the example line detection image 15 is obtained by removing the group of lines that do not contribute to the angle of the bladder neck.

[0064] The output of the results data is not limited to display on a display device. The results data can be configured to correspond to the format of reports for printing and data transmission.

[0065] As explained above, according to the structure described in this embodiment, by simply inputting an ultrasound image of the subject's pelvic region, the processing steps up to calculating the angle of the subject's bladder neck can be automated. Therefore, it is possible to provide index values ​​that establish a correlation with the function of the subject's lower urinary tract, independent of the subject's subjectivity and experience in observing the ultrasound image. In other words, it provides an environment that allows for the objective evaluation of index values ​​that establish a correlation with the function of the subject's lower urinary tract.

[0066] Another example of an indicator that can be used to evaluate the function of the aforementioned lower urinary tract is the angle of the bulbourethral region. For example... Figure 15 As shown, the bulbous urethra 3a is a portion of the urethra 3 that curves dramatically near the bladder neck 4. The angle φ of the bulbous urethra 3a is defined as the angle formed by two approximate curves extending along the two ends of the bulbous urethra 3a.

[0067] Figure 16 Examples and Figure 3 The ultrasound image 17 of the subject's urethra 3 obtained in STEP 1 (the location of the urethra 3 is highlighted by a dashed line). Since the urethra 3 is a very long and thin tubular organ, it is not easy to determine the location of the urethra 3 visually from the ultrasound image 17.

[0068] Figure 17 Will Figure 16 The ultrasound image 17 is shown upside down. That is, the urethra is projected above the bladder neck 4. The cavernous muscles 8 and soft tissue 9 are adjacent to the urethra and appear as low-brightness areas in the ultrasound image 17. Therefore, the boundary B of the high-brightness area adjacent to this low-brightness area can be regarded as the urethra 3.

[0069] Figure 18 This example illustrates the application of image processing to the image data corresponding to ultrasound image 17. Figure 3 STEP2 extracts the state of the contour of the region of interest corresponding to the low-brightness and high-brightness regions mentioned above.

[0070] In this example, YOLO is also used as the object detection algorithm. YOLO is used to extract the contours of regions whose probability of being identified as falling within a region of interest is higher than a given threshold. The contours are obtained as a set of points. When building the deep learning model that forms the basis of the YOLO used in this example, teaching data annotated with expert-labeled regions of interest is used to depict the contours.

[0071] Next, the angle of the bulbous urethra, which serves as an indicator value, is calculated. Figure 3(STEP 3). Specifically, in order to cut out the portion corresponding to the bulbous urethra 3a from the extracted contour, a first endpoint P1 and a second endpoint P2 are determined. The first endpoint P1 can be determined, for example, as the point where the above-mentioned group of points is considered as a curve. The second endpoint P2 can be determined, for example, using the coordinates of a point determined as the vertex of the bladder neck 4.

[0072] Figure 19 A first approximate straight line L11 and a second approximate straight line L12 are defined for the portion corresponding to the cut-out bulbous urethra 3a. The first approximate straight line L11 is defined by applying linear regression to a group of points within a suitable range including the first endpoint P1. The second approximate straight line L12 is defined by applying linear regression to a group of points within a suitable range including the second endpoint P2. The angle φ formed by the first approximate straight line L11 and the second approximate straight line L12 is determined as the angle of the bulbous urethra.

[0073] Next, the output of the result data is processed. Figure 3 (STEP 4). The result data involved in this example is configured to include at least the data corresponding to the angle φ calculated as described above. The output of the result data, in addition to the reference... Figure 14 In addition to or in place of the display device, the explanation can also be carried out through methods such as printing reports and sending data.

[0074] Based on the structure described in this example, by simply inputting an ultrasound image of the subject's pelvic region, the processing steps up to calculating the angle of the subject's bulbourethral region can be automated. Therefore, it is possible to provide index values ​​that establish a correlation with the subject's lower urinary tract function without relying on the subject's subjectivity or experience in observing the ultrasound image. In other words, it provides an environment that allows for the objective evaluation of index values ​​that establish a correlation with the subject's lower urinary tract function.

[0075] Figure 20 The pelvic floor muscle 5 shown is one of the multiple muscles that make up the pelvic floor. The pelvic floor is a collective term for the muscles, fascia, and ligaments that support the pelvic organs (uterus, bladder, rectum, etc.). Training to strengthen the pelvic floor muscles 5 aims to inhibit incontinence caused by decreased pelvic floor function after childbirth, and to restore lower urinary tract function that may be associated with the aforementioned total prostatectomy. In the following explanation, this training will be referred to as "pelvic floor muscle training."

[0076] Pelvic floor muscle training involves the subject consciously contracting the pelvic floor muscles 5. Contracting the pelvic floor muscles 5 narrows the distance D between the lower pubic bone 6 and the anorectal angle 7. The ability of the pelvic floor muscles 5 to contract appropriately is evaluated by a medical professional visually confirming the distance D in an ultrasound image of the subject's pelvic region. In other words, distance D serves as a quantitative indicator of pelvic floor function.

[0077] refer to Figure 3 The image evaluation method described can also be used to calculate this index value. That is, the distance D between the lower pubic margin 6 and the anorectal angle 7 can be automatically calculated based on the ultrasound image of the subject's pelvic region.

[0078] Figure 21 Examples and Figure 3 The image data obtained in STEP 1 corresponds to the ultrasound image 18 of the subject's pelvic region. Image processing is then applied to this image data. Figure 3 STEP 2) extracts the region containing at least the lower pubic margin and at least the anorectal angle as the region of interest.

[0079] In this example, YOLO is also used as the object detection algorithm. YOLO identifies regions of interest A1 as having a probability higher than a given threshold of being mapped to the subpubic margin. Similarly, YOLO identifies regions of interest A2 as having a probability higher than a given threshold of being mapped to the anorectal angle. Teaching data, which forms the basis of the deep learning model used in this example, is used to annotate the locations of the subpubic margin and anorectal angle on ultrasound images of the pelvic region.

[0080] Next, the distance D, which serves as the index value, is calculated. Figure 3 (STEP 3). In this example, the distance D is defined as the length of the straight line connecting the center point of region of interest A1 and the center point of region of interest A2. However, it is also possible to determine appropriate representative points for each region of interest and define the distance D as the length of the straight line connecting these representative points to each other.

[0081] Next, the output of the result data is processed. Figure 3 (STEP 4). The result data involved in this example consists of data corresponding to the distance D calculated as described above. The output of the result data, in addition to the reference... Figure 14 In addition to or in place of the display device, the explanation can also be carried out through methods such as printing reports and sending data.

[0082] Based on the structure described in this example, by simply inputting an ultrasound image of the subject's pelvic region, the processing steps up to calculating the distance between the subject's lower pubic bone and the anorectal angle can be automated. Therefore, it is possible to provide index values ​​that establish a correlation with the function of the subject's pelvic floor, independent of the subject's subjectivity and experience in observing the ultrasound image. In other words, it provides an environment that allows for the objective evaluation of index values ​​that establish a correlation with the function of the subject's pelvic floor.

[0083] Furthermore, as mentioned above, pelvic floor muscle training also aims to restore the function of the lower urinary tract. Therefore, the distance between the lower pubic bone and the anorectal angle can also serve as an indicator for establishing a correlation with the function of the lower urinary tract.

[0084] When the aforementioned ultrasound images 18 are acquired continuously over time, the temporal variation of distance D can be automatically obtained by applying a process that tracks the positions of the inferior pubic margin and the anorectal angle determined through image processing. Examples of this processing include KCF (Kernelized Correction Filter), MIL (Multiple Instance Learning), TLD (Tracking Learning Detection), boosting, and Median-Flow.

[0085] By obtaining the time-varying changes of distance D, we can quantitatively evaluate the contraction speed, contraction acceleration, contraction frequency, duration of contraction, ability to repeat contraction and extension, and the delay time of muscle contraction and extension commands (the self-initiation of muscle contraction).

[0086] In addition, through such Figure 21 As illustrated, displaying ultrasound images of the area of ​​interest on a display device in real time provides a clearer visualization of the contraction and relaxation of the pelvic floor muscles. This allows for better understanding of the success or failure of the exercise and the need for adjustments based on feedback from those undergoing pelvic floor muscle training.

[0087] Figure 3 The output data in STEP4 can include an estimate of the severity of the condition related to a decrease in at least one of the lower urinary tract function and pelvic floor function, based on various indicators calculated as described above.

[0088] As one example, since a greater degree of funneling corresponds to a larger angle θ at the bladder neck, the severity of lower urinary tract dysfunction is presumed to be higher. As another example, a smaller change in the distance between the lower pubic margin and the anorectal angle during pelvic floor muscle training is presumed to indicate a higher degree of pelvic floor dysfunction.

[0089] Data showing the correlation between indicator values ​​and severity of illness is pre-stored in a storage device as a tabular format or as the result of machine learning. This data is used as input to a processing device to automatically estimate the severity of illness.

[0090] The output data is structured to visualize the estimated severity of illness using at least one of text, color, and symbols.

[0091] Based on this structure, the process of estimating the severity of functional impairment in at least one of the lower urinary tract and pelvic floor based on ultrasound images projected onto the subject's pelvic region can be automated, providing an observer-based, experience-based evaluation environment independent of ultrasound images.

[0092] Figure 3 The output data in STEP4 can include the selection results of pelvic floor muscle training methods based on at least one of the various indicator values ​​calculated as described above and the estimated severity of the illness as described above.

[0093] The guidelines for selecting training methods based on the evaluation results of pelvic floor muscle contraction ability are generalized. Therefore, by defining the correspondence between at least one of the obtained index value and the estimated severity of the condition, as described above, and the evaluation results of pelvic floor muscle contraction ability, the selection of a suitable training method for the subject can be automated. Data representing the correspondence is pre-stored in a storage device as a table or as the result of machine learning. The training method is automatically selected by referring to this data by a processing device that receives at least one of the index value and the estimated severity of the condition as input.

[0094] The output data is structured to visualize the selection of the training method using at least one of text, color, and symbols.

[0095] Based on the structure, the process of selecting a suitable pelvic floor muscle training method for a subject based on ultrasound images of the subject's pelvic region can be automated, providing a guidance environment that is independent of the observer's subjectivity and experience, and does not rely on ultrasound images.

[0096] like Figure 3 As illustrated, the image evaluation method disclosed herein can be configured to compare index values ​​calculated at multiple time points with each other.

[0097] Specifically, after outputting the result data at a certain point in time, a judgment is made as to whether the indicator values ​​need to be compared (STEP5). If it is determined that no comparison is needed ("No" in STEP5), then the processing involved in this method ends.

[0098] If it is determined that a comparison of indicator values ​​is required ("Yes" in STEP5), then it is determined whether the additional indicator values ​​for comparison have been obtained (STEP6). If it is determined that additional indicator values ​​have been obtained ("No" in STEP6), then the process returns to STEP1. The process for obtaining additional indicator values ​​is repeated.

[0099] If it is determined that an additional indicator value has been obtained ("Yes" in STEP6), a comparison is made with the previously obtained indicator value, and data representing the comparison result is output (STEP7). The comparison result can be displayed by listing indicator values ​​obtained at multiple time points, or by displaying a new indicator value obtained based on multiple indicator values. Examples of new indicator values ​​can include differences, averages, etc.

[0100] Based on this structure, the observation of the function of at least one of the subject's lower urinary tract and pelvic floor, as well as the evaluation of the results of pelvic floor muscle training, can be automated. This provides an evaluation environment that is independent of the observer's subjective and experiential factors, and does not rely on ultrasound images.

[0101] Figure 3 The illustrated image evaluation method can be executed by at least one computing device. Each processing step can be executed by a separate computing device, or a single computing device can execute multiple processing steps. When the image evaluation method is executed by multiple computing devices, these devices can be located in geographically dispersed locations. In this case, data transmission between the computing devices is performed via a public or private communication network. This communication network can be implemented through wired connections, wireless connections, or combinations thereof.

[0102] Figure 22 The functional structure of the image evaluation device 30 according to one embodiment is illustrated. The image evaluation device 30 is an example of the aforementioned computing device. The image evaluation device 30 can be a device installed in a specific location or a device that can be carried by a user.

[0103] The image evaluation device 30 includes an input interface 31, a processor 32, and an output interface 33.

[0104] Input interface 31 is configured as a hardware interface for receiving image data IM from input device 40. The image data IM corresponds to an ultrasound image projected onto the pelvic region of the subject. Input device 40 can be an ultrasound probe that contacts the subject's body or a device that visualizes the ultrasound image. The image data IM corresponding to input device 40 can be in analog or digital data format. When the image data IM is in analog data format, input interface 31 includes suitable conversion circuitry including an A / D converter.

[0105] Processor 32 is configured to perform reference operations based on image data IM received from input interface 31. Figure 3The processes described herein. A processor 32 with such functionality can be implemented by a general-purpose microprocessor that operates in cooperation with general-purpose memory. Examples of general-purpose microprocessors include CPUs, MPUs, and GPUs. Examples of general-purpose memory include ROMs and RAMs. In this case, a computer program for implementing the function can be stored in the ROM. The general-purpose microprocessor can specify at least a portion of the program stored in the ROM and expand it in the RAM, cooperating with the RAM to execute the aforementioned processes. In this case, the general-purpose memory is an example of a non-transitory computer-readable medium storing a computer program.

[0106] Processor 32 can also be implemented using a microcontroller, ASIC, FPGA, or other application-specific integrated circuit (ASIC) with a storage element pre-installed with a computer program for implementing this function. In this case, the storage element is an example of a non-transitory computer-readable medium storing a computer program. Processor 32 can be implemented using a combination of a general-purpose microprocessor and an ASIC.

[0107] Output interface 33 is configured as a hardware interface for outputting the result data RS generated by processor 32 to output device 50. Examples of output device 50 include display devices, speakers, printers, data transmission devices, etc. Output device 50 can be part of image evaluation device 30 or a device independent of image evaluation device 30. The result data RS corresponds to the specifications of output device 50 and can be in analog data form or digital data form. When the result data RS is in analog data form, output interface 33 has suitable conversion circuitry including a D / A converter.

[0108] The structures described so far are merely illustrative examples to facilitate understanding of this disclosure. Each structural example can be appropriately modified and combined with other structural examples without departing from the spirit of this disclosure.

[0109] The indices that establish a correlation with at least one of the lower urinary tract function and pelvic floor function in the subject are not limited to the angle of the bladder neck, the distance between the lower pubic margin and the anorectal angle, or the angle of the bulbourethral region. Other indices can also be included in the calculation as long as they can be determined based on ultrasound images of organs contained within the subject's pelvic region and established a correlation with at least one of the lower urinary tract function and pelvic floor function. Examples of such organs include the bladder, uterus, urethra, and vagina. Examples of such indices include the distance of movement of the bladder neck and the contraction rate of the levator hiatus area.

[0110] As part of this disclosure, reference is made to Japanese Patent Application No. 2023-090404, filed on May 31, 2023.

Claims

1. An image evaluation method, which is an image evaluation method executed by at least one computing device. Obtain image data corresponding to the ultrasound image of the subject's pelvic region. Image processing is applied to the image data to extract multiple feature points related to a given location. Based on the aforementioned feature points, an index value is calculated that is associated with at least one of the lower urinary tract function and pelvic floor function.

2. The image evaluation method according to claim 1, wherein, Each of the plurality of feature points corresponds to a portion of the bladder neck. The index value corresponds to the angle of the bladder neck.

3. The image evaluation method according to claim 2, wherein, Each of the plurality of feature points is a portion of a straight line extracted by the image processing, and clustering is performed with regard to the slope of that straight line.

4. The image evaluation method according to claim 1, wherein, The aforementioned feature points correspond to the lower edge of the pubic bone and the anorectal angle. The index value corresponds to at least one of the distance between the lower edge of the pubis and the anorectal angle, and the temporal variation of that distance.

5. The image evaluation method according to claim 4, wherein, The index value is calculated by applying image tracking technology to the multiple feature points.

6. The image evaluation method according to any one of claims 1 to 5, wherein, The severity of the condition is estimated based on the aforementioned index values, specifically the reduction in at least one of the lower urinary tract function and pelvic floor function in the subject. Output the data corresponding to the estimated result.

7. The image evaluation method according to any one of claims 1 to 6, wherein, The selection of a suitable pelvic floor muscle training method for the individual is based on the aforementioned index values. Output the data corresponding to the result of the selection.

8. The image evaluation method according to any one of claims 1 to 7, wherein, Compare the values ​​of the indicator at multiple time points. Output the data corresponding to the comparison result.

9. An image evaluation device, comprising: An interface that accepts image data corresponding to an ultrasound image of the pelvic region of the object being mapped; and The processor extracts multiple feature points related to a given location by applying image processing to the image data, and calculates an index value based on the multiple feature points that is associated with at least one of lower urinary tract function and pelvic floor function.

10. A computer program that can be executed by a processor mounted on an image evaluation device. By executing the computer program, the image evaluation device performs the following processing: Receive image data corresponding to ultrasound images of the pelvic region of the subject. Image processing is applied to the image data to extract multiple feature points related to a given location. Based on the aforementioned feature points, an index value is calculated that is associated with at least one of the lower urinary tract function and pelvic floor function.

11. A non-transitory computer-readable medium storing the computer program of claim 10.

Images