An automatic organ segmentation and recognition system

Abstract

The present invention relates to the field of medical image processing, and in particular to an automatic organ segmentation and recognition system. The present invention is provided with an image acquisition module for acquiring organ image information transmitted by an acquisition terminal; a feature analysis module, which is connected to the image acquisition module, for dividing the organ image information into a plurality of image regions, extracting texture distribution features of each image region, analyzing feature distinction characterization coefficients of each image region based on texture distribution features, and calibrating labels of each image region according to the feature distinction characterization coefficients; and a segmentation and recognition module, which is respectively connected to the image acquisition module and the feature analysis module, for analyzing each image region based on the labels of each image region, adaptively identifying organ categories, and thereby improving recognition accuracy under the premise of ensuring data reliability.

Description

Technical Field

[0001] The invention relates to the field of medical image processing, and in particular to an automatic organ segmentation and recognition system. Background Art

[0002] With the development of digital technology, medical imaging has completed the transformation from physical media to digital form, which not only improves the imaging quality, but also changes the way data is stored and shared, providing convenience for organ segmentation technology. Identifying and distinguishing different organs and tissues through segmentation technology can help doctors diagnose diseases more accurately.

[0003] Chinese patent publication number: CN105389813A, discloses a method for identifying organs in medical images, including: obtaining a medical image to be processed, splitting the medical image into several two-dimensional images in the X, Y and Z axis directions respectively, and setting a detection window according to the size of the target organ; using the detection window to traverse the two-dimensional image according to the set detection step length, and obtaining the detection results in the X, Y and Z axis directions; fusing the detection results, retaining the pixels detected as positive in the three directions of X, Y and Z, so as to determine the boundary of the target organ.

[0004] However, the prior art still has the following problems: in the process of segmenting and identifying organs, the texture distribution of the connection between organs and the texture generated by the lesions is relatively discrete and irregular, which may interfere with the recognition results and reduce the accuracy of recognition. Summary of the invention

[0005] To this end, the present invention provides an automatic organ segmentation and recognition system to overcome the problem that during the organ segmentation and recognition process, the texture distribution at the connection between organs and the texture generated by lesions are relatively discrete and irregular, which may interfere with the recognition results and reduce the accuracy of recognition.

[0006] To achieve the above object, the present invention provides an automatic organ segmentation and recognition system, which comprises:

[0007] An image acquisition module, which is used to acquire organ image information transmitted by an acquisition terminal;

[0008] A feature analysis module, connected to the image acquisition module, for dividing the organ image information into a plurality of image regions, extracting texture distribution features of each of the image regions, analyzing feature distinguishing characterization coefficients of each of the image regions based on texture distribution features, and calibrating labels of each of the image regions according to the feature distinguishing characterization coefficients;

[0009] The segmentation and recognition module is connected to the image acquisition module and the feature analysis module respectively, and is used to analyze each image area based on the label of each image area and identify the organ category, including:

[0010] Based on the relative position relationship of the textures in the image area, the textures in the image area are screened, the screened textures are hidden to obtain a hidden image for the image area, the hidden image is placed in a segmentation model for analysis to obtain an organ segmentation result, and a verification operation is performed on the organ segmentation result, and whether the organ segmentation result is valid is determined based on the result of the verification operation;

[0011] Alternatively, the image region is placed into a segmentation model for analysis to obtain an organ segmentation result.

[0012] Furthermore, the feature analysis module is used to extract the texture distribution features of the image area, including:

[0013] It is used to binarize the image area, obtain the grayscale image, identify several clusters through image clustering, and use the clusters as textures;

[0014] It is used to determine the number of textures, determine the distance between each texture and adjacent textures, and determine the variance of each distance as the texture distance dispersion;

[0015] The distance between textures is determined based on the midpoint of the texture.

[0016] Furthermore, the feature analysis module is used to analyze the feature distinguishing characterization coefficients for each of the image regions, including:

[0017] Calculating the ratio of the texture quantity to the texture quantity threshold and assigning a corresponding weight value as the first characterization factor;

[0018] The ratio of the texture distance dispersion to the texture distance dispersion threshold is calculated and a corresponding weight value is assigned as a second characterization factor;

[0019] The sum of the first characterization factor and the second characterization factor is determined as a feature distinguishing characterization coefficient of the image region.

[0020] Furthermore, the feature analysis module is used to mark the labels of each image area, including:

[0021] If the feature discrimination representation coefficient is greater than or equal to the feature discrimination representation coefficient threshold, the image area is marked as a discrete distribution texture label;

[0022] If the feature discrimination representation coefficient is less than the feature discrimination representation coefficient threshold, the image area is marked as a regularly distributed texture label.

[0023] Furthermore, the segmentation and recognition module is used to identify organ types including:

[0024] If the label of the image region is a discrete distribution texture label, based on the relative position relationship of the textures in the image region, the textures in the image region are screened, the screened textures are hidden to obtain a hidden image for the image region, the hidden image is placed in a segmentation model for analysis to obtain an organ segmentation result, and a verification operation is performed on the organ segmentation result, and whether the organ segmentation result is valid is determined based on the result of the verification operation;

[0025] If the label of the image region is a regularly distributed texture label, the image region is placed in the segmentation model for analysis to obtain the organ segmentation result.

[0026] Furthermore, the segmentation and recognition module is used to screen the textures in the image area based on the relative position relationship of the textures in the image area, including:

[0027] To determine the direction reference line of each texture, determine the direction reference line corresponding to each texture in each image area and the angle between the direction reference lines;

[0028] The image region is divided into a plurality of image sub-regions, and classified into feature sub-regions and non-feature sub-regions based on the angles and the number of directional reference lines in the image sub-regions, and the textures in the non-feature sub-regions are screened out;

[0029] The characteristic sub-regions must satisfy that the angle between any two texture corresponding direction reference lines in the characteristic sub-region is less than a predetermined angle threshold and the number of characteristic sub-regions must be greater than a predetermined number threshold.

[0030] Furthermore, the segmentation and recognition module is used to determine the texture corresponding direction reference line including:

[0031] It is used to identify the distance between each pixel point in the texture, and connect the pixel points with the maximum distance to obtain the direction reference line.

[0032] Furthermore, the segmentation recognition module is used to verify the organ segmentation result, including:

[0033] Used to restore the filter texture;

[0034] Used to place the image in the image area into the segmentation model to obtain a number of output results;

[0035] It is used to determine whether each of the output results is consistent with the output result of the segmentation model for the hidden image, and determine the consistency ratio.

[0036] Furthermore, the segmentation recognition module is used to determine whether the organ segmentation result is valid based on the result of the verification operation, including:

[0037] If the consistent ratio is greater than a predetermined ratio threshold, the organ segmentation result is determined to be valid.

[0038] Furthermore, it also includes a display output module, which is connected to the segmentation and recognition module and is used to display corresponding content based on the organ segmentation result.

[0039] Compared with the prior art, the present invention is provided with an image acquisition module for acquiring organ image information transmitted by an acquisition terminal; a feature analysis module, which is connected to the image acquisition module, for dividing the organ image information into a number of image areas, extracting texture distribution features of each image area, analyzing feature distinction characterization coefficients of each image area based on texture distribution features, and calibrating labels of each image area according to the feature distinction characterization coefficients; a segmentation and recognition module, which is respectively connected to the image acquisition module and the feature analysis module, for analyzing each image area based on the label of each image area, adaptively identifying organ categories, and thereby improving recognition accuracy while ensuring data reliability.

[0040] In particular, the present invention analyzes the feature distinguishing characterization coefficient of each image area based on the texture distribution feature. In the process of organ segmentation and recognition, the edge contour of the organ is mostly referred to. In actual situations, the boundaries of some edge contours of organs are relatively clear, but due to differences in the human body, the interweaving of tissues or diseased tissues, some edge contours are unclear, and there are many relatively discrete and irregular texture segments distributed around them. These areas may be the junction areas of a single organ and other organs, and have certain data representativeness, but the texture is complex and certain noise is introduced, which affects the organ segmentation judgment. Therefore, the present invention calculates the feature distinguishing characterization coefficient of each image area through a feature analysis module, and distinguishes the labels of each image area, so that the segmentation and recognition module can adaptively select the analysis method for different image areas according to the labels, thereby improving the accuracy of organ segmentation while ensuring data reliability.

[0041] In particular, when the label of the image area is a discrete distribution texture label, the present invention screens the texture in the image area based on the relative position relationship of the texture in the image area. In actual situations, some image areas under discrete distribution texture labels have too much texture, which will hide the real organ contour. However, due to the growth of the organ's own tissue, the texture around the real organ contour will spread along the organ contour. Even if the real organ contour is unclear and there are many breakpoints, the spread trend of most textures has a certain directionality. Therefore, the present invention considers constructing a directional reference line of the texture for comparison to determine the spread trend of the texture, and removes relatively discrete non-feature sub-areas together, and screens and hides the interfering texture, reducing interference information, and thereby improving the accuracy of organ segmentation.

[0042] In particular, when the label of the image area is a discretely distributed texture label, the present invention performs a verification operation on the organ segmentation result after obtaining it to verify whether it will cause a large difference in the organ segmentation result, determine whether the organ segmentation result is valid, and thus improve the recognition accuracy while ensuring data reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a functional module diagram of the automatic organ segmentation and recognition system of an embodiment of the invention;

[0044] Figure 2 A logical decision diagram for labeling each image region for an embodiment of the invention;

[0045] Figure 3 A logical decision diagram for identifying organ classes for embodiments of the invention;

[0046] Figure 4 This is a logical decision diagram for determining whether the organ segmentation result is valid according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0048] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0049] It should be noted that, in the description of the present invention, unless otherwise clearly specified and limited, the term "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] See also Figures 1 to 4 As shown, Figure 1 This is a functional module diagram of the automatic organ segmentation and recognition system of an embodiment of the invention. Figure 2 A logical decision diagram for calibrating labels of each image region for an embodiment of the invention, Figure 3 A logical decision diagram for identifying organ classes for an embodiment of the invention, Figure 4 The present invention is a logic decision diagram for determining whether the organ segmentation result is valid. The automatic organ segmentation and recognition system of the present invention comprises:

[0051] An image acquisition module, which is used to acquire organ image information transmitted by an acquisition terminal;

[0052] A feature analysis module, connected to the image acquisition module, for dividing the organ image information into a plurality of image regions, extracting texture distribution features of each of the image regions, analyzing feature distinguishing characterization coefficients of each of the image regions based on texture distribution features, and calibrating labels of each of the image regions according to the feature distinguishing characterization coefficients;

[0053] The segmentation and recognition module is connected to the image acquisition module and the feature analysis module respectively, and is used to analyze each image area based on the label of each image area and identify the organ category, including:

[0054] Based on the relative position relationship of the textures in the image area, the textures in the image area are screened, the screened textures are hidden to obtain a hidden image for the image area, the hidden image is placed in a segmentation model for analysis to obtain an organ segmentation result, and a verification operation is performed on the organ segmentation result, and whether the organ segmentation result is valid is determined based on the result of the verification operation;

[0055] Alternatively, the image region is placed into a segmentation model for analysis to obtain an organ segmentation result.

[0056] Specifically, there is no specific limitation on the type of acquisition terminal. In this embodiment, organ image information can be acquired through an X-ray scanning device, which will not be described in detail.

[0057] Specifically, there is no specific limitation on the segmentation model. Those skilled in the art can use an open source organ segmentation model to achieve segmentation and recognition of organs in organ images. Of course, they can also build their own training sets and sample sets to train the segmentation model, which will not be repeated here.

[0058] Specifically, there is no limitation on the specific structure of the feature analysis module and the segmentation and recognition module, and the module itself or each unit therein can be composed of logic components or a combination of logic components, and the logic components include a field programmable processor, a computer or a microprocessor in a computer.

[0059] Specifically, the feature analysis module is used to extract the texture distribution features of the image area, including:

[0060] It is used to binarize the image area, obtain the grayscale image, identify several clusters through image clustering, and use the clusters as textures;

[0061] It is used to determine the number of textures, determine the distance between each texture and adjacent textures, and determine the variance of each distance as the texture distance dispersion;

[0062] The distance between textures is determined based on the midpoints of the textures, and usually the distance between the corresponding midpoints of the textures is determined as the distance between the textures.

[0063] It is understandable that, since texture is usually significantly different from background, relevant textures can be identified by clustering clusters, which will not be described in detail.

[0064] Specifically, the feature analysis module is used to analyze the feature distinguishing characterization coefficients for each of the image regions, including:

[0065] Calculating the ratio of the number of textures to the threshold value of the number of textures and assigning a corresponding weight value α as the first characterization factor;

[0066] The ratio of the texture distance dispersion to the texture distance dispersion threshold is calculated and a corresponding weight value β is assigned as the second characterization factor;

[0067] The sum of the first characterization factor and the second characterization factor is determined as a feature distinguishing characterization coefficient of the image region.

[0068] Specifically, the feature discrimination representation coefficient of each image region is calculated according to the following formula:

[0069]

[0070] Where F represents the feature discrimination coefficient, N represents the number of textures, N0 represents the texture number threshold, L represents the texture distance discreteness, L0 represents the texture distance discreteness threshold, α represents the texture number weight coefficient, and β represents the texture distance discreteness weight coefficient.

[0071] Texture distance dispersion is the variance of each texture distance;

[0072] In this embodiment, the texture quantity threshold N0 and the texture distance discreteness threshold L0 are obtained in advance, and a number of organ image data are obtained, the texture quantity and the texture distance discreteness are extracted, the texture quantity mean ΔN and the texture distance discreteness mean ΔL are solved, and N0=r1×ΔN, L0=r2×ΔL are set, wherein r1 is the first deviation coefficient, r2 is the second deviation coefficient, 1.05<r1<1.1, 1.1<r2<1.15, α is 0.39, and β is 0.61.

[0073] The present invention analyzes the feature distinguishing characterization coefficient of each image area based on texture distribution features. In the process of organ segmentation and recognition, the edge contour of the organ is mostly referred to. In actual situations, the boundaries of some edge contours of organs are relatively clear, but due to differences in the human body, the interweaving of tissues or diseased tissues, some edge contours are unclear, and there are many relatively discrete and irregular texture segments distributed around them. These areas may be the junction areas of a single organ and other organs, and have certain data representativeness, but the texture is complex and certain noise is introduced, which affects the organ segmentation judgment. Therefore, the present invention calculates the feature distinguishing characterization coefficient of each image area through a feature analysis module, distinguishes the labels of each image area, so that the segmentation and recognition module can adaptively select the analysis method for different image areas according to the labels, thereby improving the accuracy of organ segmentation while ensuring data reliability.

[0074] Specifically, the feature analysis module is used to mark the labels of each image area, including:

[0075] If the feature discrimination representation coefficient is greater than or equal to the feature discrimination representation coefficient threshold, the image area is marked as a discrete distribution texture label;

[0076] If the feature discrimination representation coefficient is less than the feature discrimination representation coefficient threshold, the image area is marked as a regularly distributed texture label.

[0077] The feature discrimination characterization coefficient F0 is selected in the interval [1.27, 1.38].

[0078] Specifically, the segmentation and recognition module is used to identify organ types including:

[0079] If the label of the image region is a discrete distribution texture label, based on the relative position relationship of the textures in the image region, the textures in the image region are screened, the screened textures are hidden to obtain a hidden image for the image region, the hidden image is placed in a segmentation model for analysis to obtain an organ segmentation result, and a verification operation is performed on the organ segmentation result, and whether the organ segmentation result is valid is determined based on the result of the verification operation;

[0080] If the label of the image region is a regularly distributed texture label, the image region is placed in the segmentation model for analysis to obtain the organ segmentation result.

[0081] Specifically, the segmentation and recognition module is used to screen the textures in the image area based on the relative position relationship of the textures in the image area, including:

[0082] To determine the direction reference line of each texture, determine the direction reference line corresponding to each texture in each image area and the angle between the direction reference lines;

[0083] The image region is divided into a plurality of image sub-regions, and classified into feature sub-regions and non-feature sub-regions based on the angles and the number of directional reference lines in the image sub-regions, and the textures in the non-feature sub-regions are screened out;

[0084] The characteristic sub-regions must satisfy that the angle between any two texture corresponding direction reference lines in the characteristic sub-region is less than a predetermined angle threshold and the number of characteristic sub-regions must be greater than a predetermined number threshold.

[0085] In this embodiment, the predetermined angle threshold J0 and the predetermined quantity threshold Z0 are obtained by pre-determination, and image data of a number of organs are obtained. The angle mean ΔJ and the quantity mean ΔZ between the texture direction reference lines in each image area are extracted, and J0=c1×ΔJ and Z0=c2×ΔZ are set, wherein c1 is the first offset coefficient, c2 is the second offset coefficient, 1.5<c1<1.8, 1.3<c2<1.5.

[0086] Specifically, the segmentation and recognition module is used to determine the texture corresponding direction reference line including:

[0087] It is used to identify the distance between each pixel point in the texture, and connect the pixel points with the maximum distance to obtain the direction reference line.

[0088] It is understandable that in actual situations, organ textures are mostly distributed in strips, so the pixels with the maximum distance are used to determine the two ends of the texture, and the direction reference line is used to represent the spreading direction of the texture.

[0089] In the case where the label of the image area is a discrete distribution texture label, the present invention screens the texture in the image area based on the relative position relationship of the texture in the image area. In actual situations, the image area under some discrete distribution texture labels has too much texture, which will hide the real organ contour. However, due to the growth of the organ's own tissue, the texture around the real organ contour will spread along the organ contour. Even if the real organ contour is not clear and there are many breakpoints, the spreading trend of most textures has a certain directionality. Therefore, the present invention considers constructing a directional reference line of the texture for comparison to determine the spreading trend of the texture, and relatively discrete non-feature sub-areas are eliminated together, and the interfering texture is screened and hidden, so as to reduce the interference information and thereby improve the accuracy of organ segmentation.

[0090] Specifically, the segmentation recognition module is used to verify the organ segmentation result, including:

[0091] Used to restore the filter texture;

[0092] Used to place the image in the image area into the segmentation model to obtain a number of output results;

[0093] It is used to determine whether each of the output results is consistent with the output result of the segmentation model for the hidden image, and determine the consistency ratio.

[0094] It can be understood that each of the output results has a one-to-one correspondence with the output result of the segmentation model for the hidden image, and the consistency ratio is the ratio of the number of consistent output results to the total number of output results, which will not be elaborated here.

[0095] It can be understood that the output result is the organ segmentation result, and the consistent output result means that the organ segmentation and recognition results are consistent.

[0096] In the case where the labels of the image areas are discretely distributed texture labels, the present invention performs a verification operation on the organ segmentation results after obtaining them to verify whether the verification will cause large differences in the organ segmentation results and determine whether the organ segmentation results are valid, thereby improving the recognition accuracy while ensuring data reliability.

[0097] Specifically, the segmentation recognition module is used to determine whether the organ segmentation result is valid based on the result of the verification operation, including:

[0098] If the consistent ratio is greater than a predetermined ratio threshold, the organ segmentation result is determined to be valid;

[0099] If the consistent ratio is less than or equal to the predetermined ratio threshold, the organ segmentation result is determined to be invalid.

[0100] Specifically, in order to reflect the accuracy of the organ segmentation results, the ratio threshold is selected in the interval [80%, 90%].

[0101] Specifically, it also includes a display output module, which is connected to the segmentation and recognition module and is used to display corresponding content based on the organ segmentation result.

[0102] Specifically, there is no limitation on the specific structure of the display output module, which only needs to have the function of displaying the corresponding content of the organ segmentation result, for example, an electronic display screen, which will not be elaborated here.

[0103] It is understandable that the corresponding content of the organ segmentation result includes organ name and organ morphology.

[0104] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. An automatic organ segmentation and recognition system, characterized in that: include: An image acquisition module, which is used to acquire organ image information transmitted by an acquisition terminal; A feature analysis module, connected to the image acquisition module, for dividing the organ image information into a plurality of image regions, extracting texture distribution features of each of the image regions, analyzing feature distinguishing characterization coefficients of each of the image regions based on texture distribution features, and calibrating labels of each of the image regions according to the feature distinguishing characterization coefficients; The segmentation and recognition module is connected to the image acquisition module and the feature analysis module respectively, and is used to analyze each image area based on the label of each image area and identify the organ category, including: Based on the relative position relationship of the textures in the image area, the textures in the image area are screened, the screened textures are hidden to obtain a hidden image for the image area, the hidden image is placed in a segmentation model for analysis to obtain an organ segmentation result, and a verification operation is performed on the organ segmentation result, and whether the organ segmentation result is valid is determined based on the result of the verification operation; Or, the image region is placed into the segmentation model for analysis to obtain the organ segmentation result; The feature analysis module is used to analyze the feature distinguishing characterization coefficients for each of the image regions, including: Calculating the ratio of the texture quantity to the texture quantity threshold and assigning a corresponding weight value as the first characterization factor; The ratio of the texture distance dispersion to the texture distance dispersion threshold is calculated and a corresponding weight value is assigned as a second characterization factor; for determining the sum of the first characterization factor and the second characterization factor as a characteristic distinguishing characterization coefficient of the image region; The segmentation and recognition module is used to recognize organ types, including: If the label of the image region is a discrete distribution texture label, based on the relative position relationship of the textures in the image region, the textures in the image region are screened, the screened textures are hidden to obtain a hidden image for the image region, the hidden image is placed in a segmentation model for analysis to obtain an organ segmentation result, and a verification operation is performed on the organ segmentation result, and whether the organ segmentation result is valid is determined based on the result of the verification operation; If the label of the image region is a regularly distributed texture label, the image region is placed in the segmentation model for analysis to obtain the organ segmentation result.

2. The automatic organ segmentation and recognition system according to claim 1, characterized in that: The feature analysis module is used to extract the texture distribution features of the image area, including: It is used to binarize the image area, obtain the grayscale image, identify several clusters through image clustering, and use the clusters as textures; It is used to determine the number of textures, determine the distance between each texture and adjacent textures, and determine the variance of each distance as the texture distance dispersion; The distance between textures is determined based on the midpoint of the texture.

3. The automatic organ segmentation and recognition system according to claim 1, characterized in that: The feature analysis module is used to mark the labels of each image area, including: If the feature discrimination representation coefficient is greater than or equal to the feature discrimination representation coefficient threshold, the image area is marked as a discrete distribution texture label; If the feature discrimination representation coefficient is less than the feature discrimination representation coefficient threshold, the image area is marked as a regularly distributed texture label.

4. The automatic organ segmentation and recognition system according to claim 1, characterized in that: The segmentation and recognition module is used to screen the textures in the image area based on the relative position relationship of the textures in the image area, including: To determine the direction reference line of each texture, determine the direction reference line corresponding to each texture in each image area and the angle between the direction reference lines; The image region is divided into a plurality of image sub-regions, and classified into feature sub-regions and non-feature sub-regions based on the angles and the number of directional reference lines in the image sub-regions, and the textures in the non-feature sub-regions are screened out; The characteristic sub-regions must satisfy that the angle between any two texture corresponding direction reference lines in the characteristic sub-region is less than a predetermined angle threshold and the number of characteristic sub-regions must be greater than a predetermined number threshold.

5. The automatic organ segmentation and recognition system according to claim 4, characterized in that: The segmentation and recognition module is used to determine the texture corresponding direction reference line, including: It is used to identify the distance between each pixel point in the texture, and connect the pixel points with the maximum distance to obtain the direction reference line.

6. The automatic organ segmentation and recognition system according to claim 1, characterized in that: The segmentation recognition module is used to verify the organ segmentation result, including: Used to restore the filter texture; Used to place the image in the image area into the segmentation model to obtain a number of output results; It is used to determine whether each of the output results is consistent with the output result of the segmentation model for the hidden image, and determine the consistency ratio.

7. The automatic organ segmentation and recognition system according to claim 1, characterized in that: The segmentation recognition module is used to determine whether the organ segmentation result is valid based on the result of the verification operation, including: If the consistent ratio is greater than a predetermined ratio threshold, the organ segmentation result is determined to be valid.

8. The automatic organ segmentation and recognition system according to claim 1, characterized in that: It also includes a display output module, which is connected to the segmentation and recognition module and is used to display corresponding content based on the organ segmentation result.

Images