Lumbar disease responsibility segment determination method and related equipment

Abstract

The embodiment of the invention discloses a liability segment judgment method for lumbar vertebra diseases and related equipment, and relates to the technical field of image recognition, the method comprises the steps that to-be-recognized image data are acquired, and the to-be-recognized image data comprise a sagittal view image of the lumbar vertebra and a cross section image of the lumbar vertebra; according to the first labeling information of each segment of the lumbar vertebra and the second labeling information of the focus on the lumbar vertebra, labeling the lumbar vertebra sagittal view image to obtain a first sagittal view image with the first labeling information and a second sagittal view image with the second labeling information; according to the first sagittal view image and the second sagittal view image, identifying a first responsibility segment through a first preset YOLO model; labeling the cross section image based on the third labeling information to obtain a target cross section image; according to the target cross section image, identifying a second responsibility segment through a second preset YOLO model; and inputting the first responsibility segment and the second responsibility segment into a preset classifier to obtain a classification result, and marking the to-be-identified image data based on the classification result to complete the judgment of the responsibility segment of the lumbar vertebra disease.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and in particular to a method and related equipment for determining the responsible segment in lumbar spine diseases. Background Technology

[0002] Lumbar spine disorders, especially multi-segmental lumbar lateral recess stenosis, are common in the elderly, with main symptoms including lower back pain and lower extremity neurological symptoms. Since the degree of stenosis on imaging is not entirely correlated with clinical symptoms, accurately identifying the responsible segment (i.e., the primary segment causing symptoms) is crucial for clinical treatment. Traditional methods for determining the responsible segment rely primarily on a comprehensive analysis of imaging examinations (such as MRI and CT), clinical symptoms, and the physician's experience. However, this method has limitations, such as strong diagnostic subjectivity and reliance on invasive diagnostic techniques (such as selective nerve root blocks). Furthermore, imaging examinations themselves are affected by factors such as physician experience, image quality, and differences in machine operation. Summary of the Invention

[0003] In view of this, the present invention provides a method and related equipment for determining the responsible segment of lumbar spine diseases, which solves the limitations of existing technologies such as strong subjectivity and reliance on invasive diagnostic methods (such as selective nerve root block).

[0004] In a first aspect, embodiments of the present invention provide a method for determining the responsible segment in lumbar spine diseases, the method comprising: Acquire image data to be identified, wherein the image data to be identified includes a sagittal image of the lumbar spine and a cross-sectional image of the lumbar spine; Based on the first annotation information of each segment of the lumbar vertebra and the second annotation information of the lesion on the lumbar vertebra, the sagittal image of the lumbar vertebra is annotated to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information; Based on the first sagittal image and the second sagittal image, the first responsible segment is identified using a first preset YOLO model; The cross-sectional image is annotated based on the third annotation information to obtain the target cross-sectional image; Based on the target cross-sectional image, the second responsible segment is identified using a second preset YOLO model; The first and second responsible segments are input into a preset classifier to obtain classification results. The image data to be identified is then labeled based on the classification results to complete the determination of the responsible segment for lumbar spine diseases.

[0005] Optionally, the step of annotating the sagittal images of the lumbar vertebrae according to the first annotation information of each segment of the lumbar vertebrae and the second annotation information of the lesions on the lumbar vertebrae, to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information, includes: The category, center information, and bounding box information of each segment of the lumbar vertebra are used as the first annotation information, and the sagittal image of the lumbar vertebra is annotated according to the first annotation information of each segment of the lumbar vertebra to obtain a first sagittal image including the first annotation information. The center location information of the lesion on the lumbar spine is used as the second annotation information, and the sagittal image of the lumbar spine is annotated according to the second annotation information of the lesion on the lumbar spine to obtain a second sagittal image with the second annotation information.

[0006] Optionally, the step of identifying the first responsible segment based on the first sagittal image and the second sagittal image using a first preset YOLO model includes: The first sagittal image is input into the first recognition unit in the first preset YOLO model to identify the segments on the lumbar spine, and the segments on the lumbar spine identified by the first recognition unit are used as candidate segments. The second sagittal image is input into the second recognition unit in the first preset YOLO model to identify the lesion on the lumbar spine, and the lesion on the lumbar spine identified by the second recognition unit is taken as the target lesion; Based on the second annotation information corresponding to the target lesion and the first annotation information corresponding to the candidate segment, candidate segments that have a preset association relationship with the target lesion are determined, and the candidate segments that have a preset association relationship with the target lesion are designated as the first responsible segments.

[0007] Optionally, the step of determining candidate segments with a preset association relationship to the target lesion based on the second annotation information corresponding to the target lesion and the first annotation information corresponding to the candidate segments, and using the candidate segments with the preset association relationship to the target lesion as the first responsible segments, includes: The second annotation information corresponding to the target lesion is used as the first discrimination data, and the first annotation information corresponding to the candidate segment is used as the second discrimination data. The first discrimination data and the second discrimination data are matched to obtain the matching result. Based on the matching results, candidate segments that have a preset association with the target lesion are determined, and the candidate segments that have a preset association with the target lesion are designated as the first responsible segments.

[0008] Optionally, the step of matching the first discrimination data and the second discrimination data to determine candidate segments that have a preset association with the target lesion, and using the candidate segments that have a preset association with the target lesion as the first responsible segment, includes: When the matching result indicates that the target lesion is within any candidate segment, the candidate segment where the target lesion is located is taken as a candidate segment with a preset association relationship with the target lesion, and the candidate segment with a preset association relationship with the target lesion is taken as the first responsible segment; When the matching result indicates that the target lesion is not among all candidate segments, the candidate segment closest to the target lesion is taken as the candidate segment with a preset association relationship with the target lesion, and the candidate segment with the preset association relationship with the target lesion is taken as the first responsible segment.

[0009] Optionally, the step of annotating the cross-sectional image based on the third annotation information to obtain the target cross-sectional image includes: Obtain historical annotation information from historical cross-sectional images, and use the information category corresponding to the historical annotation information as the information category of the third annotation information; The cross-sectional images are stitched together to obtain the image to be labeled; The cross-sectional image is annotated based on the third annotation information to obtain the target cross-sectional image.

[0010] Optionally, the step of inputting the first responsibility segment and the second responsibility segment into a preset classifier to obtain a classification result includes: The first responsible segment is input into the feature extraction unit of the preset classifier to obtain the first segment feature and the first supplementary feature of the first responsible segment. The first supplementary feature includes the brightness feature of the cross-sectional image. The classification unit of the preset classifier classifies the first responsible segment according to the first segment features and the first supplementary features to obtain a first classification result; The second responsibility segment is input into the feature extraction unit of the preset classifier to obtain the second segment feature and the second supplementary feature of the second responsibility segment. The second supplementary feature includes the brightness feature of the cross-sectional image. The classification unit of the preset classifier classifies the second responsibility segment according to the second segment features and the second supplementary features to obtain a second classification result; The first classification result and the second classification result are used together as the classification result.

[0011] On the other hand, this application provides a system for determining the responsible segment in lumbar spine diseases, the system comprising: The data acquisition module is used to acquire image data to be identified, wherein the image data to be identified includes a sagittal image of the lumbar spine and a cross-sectional image of the lumbar spine; The first annotation module is used to annotate the sagittal image of the lumbar spine according to the first annotation information of each segment of the lumbar spine and the second annotation information of the lesion on the lumbar spine, so as to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information. The first identification module is used to identify the first responsible segment based on the first sagittal image and the second sagittal image using a first preset YOLO model; The second annotation module is used to annotate the cross-sectional image based on the third annotation information to obtain the target cross-sectional image; The second identification module is used to identify the second responsible segment based on the target cross-sectional image using a second preset YOLO model; The determination module is used to input the first responsible segment and the second responsible segment into a preset classifier to obtain a classification result, and to annotate the image data to be identified based on the classification result in order to complete the determination of the responsible segment of lumbar spine disease.

[0012] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described method for determining the responsible segment of lumbar spine diseases.

[0013] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for determining the responsible segment of lumbar spine diseases as described above.

[0014] The technical solution of this invention involves acquiring image data to be identified, including sagittal images and cross-sectional images of the lumbar spine. Based on first annotation information for each segment of the lumbar spine and second annotation information for lesions on the lumbar spine, the sagittal images are annotated to obtain a first sagittal image with first annotation information and a second sagittal image with second annotation information. Based on the first and second sagittal images, a first responsible segment is identified using a first preset YOLO model. The cross-sectional images are annotated based on the third annotation information to obtain a target cross-sectional image. Based on the target cross-sectional image, a second responsible segment is identified using a second preset YOLO model. The first and second responsible segments are input into a preset classifier to obtain a classification result, and the image data to be identified is annotated based on the classification result to complete the determination of the responsible segment in lumbar spine diseases. By combining sagittal and cross-sectional images of the lumbar spine for analysis, the lesion characteristics from different perspectives are fully utilized to accurately determine the location of the responsible segment in multi-segment lumbar spine diseases. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] in: Figure 1 This is a flowchart illustrating a method for determining the responsible segment in lumbar spine diseases according to an embodiment of the present invention. Figure 2 This is a sagittal view of the lumbar spine in a method for determining the responsible segment of a lumbar spine disease according to an embodiment of the present invention. Figure 3 This is a cross-sectional view of the lumbar spine in a method for determining the responsible segment of a lumbar spine disease according to an embodiment of the present invention. Figure 4 This is a first sagittal view in a method for determining the responsible segment of a lumbar spine disease according to an embodiment of the present invention; Figure 5 This is a second sagittal view in a method for determining the responsible segment of lumbar spine diseases according to an embodiment of the present invention; Figure 6 This is a target cross-sectional view in a method for determining the responsible segment of lumbar spine diseases according to an embodiment of the present invention; Figure 7 This is a diagram of the first responsible segment in a method for determining the responsible segment of lumbar spine diseases according to an embodiment of the present invention; Figure 8 This is a diagram of the second responsible segment in a method for determining the responsible segment of lumbar spine diseases according to an embodiment of the present invention; Figure 9 This is a diagram showing the result of determining the responsible segment in a method for determining the responsible segment of a lumbar spine disease according to an embodiment of the present invention. Figure 10 This is a schematic diagram of a system for determining the responsible segment of lumbar spine diseases according to an embodiment of the present invention; Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application; Figure 12 This is a schematic diagram of the structure of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] like Figure 1 As shown in the figure, a method for determining the responsible segment of lumbar spine diseases according to an embodiment of the present invention specifically includes the following steps: S110. Acquire image data to be identified, wherein the image data to be identified includes a sagittal image of the lumbar spine and a cross-sectional image of the lumbar spine; For example, such as Figure 2 As shown, a sagittal image is a medical imaging technique that slices the human body or organ along the sagittal plane (i.e., the anterior-posterior direction) to reveal the internal structure of the body or organ.

[0019] For example, such as Figure 3 As shown, a cross-sectional image is a graphic representation of an object or structure (lumbar vertebra) cut into cross-sections according to specific segments and drawn from these cross-sections.

[0020] S120. Based on the first annotation information of each segment of the lumbar vertebra and the second annotation information of the lesion on the lumbar vertebra, the sagittal image of the lumbar vertebra is annotated to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information. In one possible implementation, the step of annotating the sagittal image of the lumbar spine according to the first annotation information of each segment of the lumbar spine and the second annotation information of the lesion on the lumbar spine, to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information, includes: The category, center information, and bounding box information of each segment of the lumbar vertebra are used as the first annotation information, and the sagittal image of the lumbar vertebra is annotated according to the first annotation information of each segment of the lumbar vertebra to obtain a first sagittal image including the first annotation information. The center location information of the lesion on the lumbar spine is used as the second annotation information, and the sagittal image of the lumbar spine is annotated according to the second annotation information of the lesion on the lumbar spine to obtain a second sagittal image with the second annotation information.

[0021] For example, for sagittal slices, two independent YOLO annotation files are generated for each image: one file is used to annotate the locations of all lumbar vertebral segments, such as... Figure 4 As shown, each segment's category number (category), center point coordinates (center information), confidence level, and region width and height (bounding box information) are recorded; another file is used to label the location of the responsible segment, such as... Figure 5 As shown, only the central location information and confidence level of the lesion are included.

[0022] S130. Based on the first sagittal image and the second sagittal image, the first responsible segment is identified using a first preset YOLO model; In one possible implementation, the step of identifying the first responsible segment based on the first sagittal image and the second sagittal image using a first preset YOLO model includes: The first sagittal image is input into the first recognition unit in the first preset YOLO model to identify the segments on the lumbar spine, and the segments on the lumbar spine identified by the first recognition unit are used as candidate segments. The second sagittal image is input into the second recognition unit in the first preset YOLO model to identify the lesion on the lumbar spine, and the lesion on the lumbar spine identified by the second recognition unit is taken as the target lesion; Based on the second annotation information corresponding to the target lesion and the first annotation information corresponding to the candidate segment, candidate segments that have a preset association relationship with the target lesion are determined, and the candidate segments that have a preset association relationship with the target lesion are designated as the first responsible segments.

[0023] For example, the first identification unit in the first preset YOLO model (which can be understood as the YOLO model) identifies the location of each segment of the lumbar spine. The YOLO model is used to further locate the responsible segment in the sagittal image, and the segment on the lumbar spine identified by the first identification unit is used as a candidate segment. The second identification unit in the first preset YOLO model (which can be understood as the YOLO model) is used to further locate the lesion in the sagittal image, and the lesion on the lumbar spine identified by the first identification unit is used as the target lesion.

[0024] In one possible implementation, the step of determining candidate segments with a preset association relationship to the target lesion based on second annotation information corresponding to the target lesion and first annotation information corresponding to the candidate segments, and designating the candidate segments with the preset association relationship to the target lesion as the first responsible segments, includes: The second annotation information corresponding to the target lesion is used as the first discrimination data, and the first annotation information corresponding to the candidate segment is used as the second discrimination data. The first discrimination data and the second discrimination data are matched to obtain the matching result. Based on the matching results, candidate segments that have a preset association with the target lesion are determined, and the candidate segments that have a preset association with the target lesion are designated as the first responsible segments.

[0025] For example, the step of matching the first discrimination data and the second discrimination data to determine candidate segments that have a preset association with the target lesion, and selecting the candidate segments that have a preset association with the target lesion as the first responsible segment, includes: When the matching result indicates that the target lesion is within any candidate segment, the candidate segment where the target lesion is located is taken as a candidate segment with a preset association relationship with the target lesion, and the candidate segment with a preset association relationship with the target lesion is taken as the first responsible segment; When the matching result indicates that the target lesion is not among all candidate segments, the candidate segment closest to the target lesion is selected as the candidate segment with a preset association relationship with the target lesion, and the candidate segment with the preset association relationship with the target lesion is selected as the first responsible segment. Specifically, for each target lesion, the system checks whether the center point of the target lesion lies within the boundary of a certain lumbar vertebral segment. If a match is successful, the candidate segment containing the target lesion is designated as the first responsible segment. If no match is found, the system calculates the distance between the center point of the target lesion and the center points of each lumbar vertebral segment, selecting the lumbar vertebral segment closest to the center point of the target lesion as the first responsible segment. Figure 7 The diagram shows the first responsible segment.

[0026] S140. Annotate the cross-sectional image based on the third annotation information to obtain the target cross-sectional image; In one possible implementation, the step of annotating the cross-sectional image based on the third annotation information to obtain the target cross-sectional image includes: Obtain historical annotation information from historical cross-sectional images, and use the information category corresponding to the historical annotation information as the information category of the third annotation information; The cross-sectional images are stitched together to obtain the image to be labeled; The cross-sectional image is annotated based on the third annotation information to obtain the target cross-sectional image.

[0027] For example, historical annotation information on historical cross-sectional images can be understood as annotation information verified by doctors, that is, annotation information obtained according to the "gold standard of responsibility segment".

[0028] For example, before stitching the cross-sectional images, the cross-sectional images are cropped to extract a square region containing the lumbar spine, and irrelevant redundant information is removed. The cross-sectional images are then annotated based on the third annotation information to obtain the following: Figure 6 The target cross-sectional image.

[0029] S150. Based on the target cross-sectional image, the second responsible segment is identified using the second preset YOLO model; For example, the YOLO model is used to detect the bounding box, and the specific lumbar segment to which the target lesion in the cross-section belongs is accurately determined based on the position of the bounding box, resulting in the following: Figure 8 The diagram shows the second responsibility segment.

[0030] S160. Input the first responsible segment and the second responsible segment into a preset classifier to obtain the classification result, and label the image data to be identified based on the classification result to complete the determination of the responsible segment of lumbar spine disease.

[0031] In one possible implementation, the step of inputting the first responsibility segment and the second responsibility segment into a preset classifier to obtain a classification result includes: The first responsible segment is input into the feature extraction unit of the preset classifier to obtain the first segment feature and the first supplementary feature of the first responsible segment. The first supplementary feature includes the brightness feature of the cross-sectional image. The classification unit of the preset classifier classifies the first responsible segment according to the first segment features and the first supplementary features to obtain a first classification result; The second responsibility segment is input into the feature extraction unit of the preset classifier to obtain the second segment feature and the second supplementary feature of the second responsibility segment. The second supplementary feature includes the brightness feature of the cross-sectional image. The classification unit of the preset classifier classifies the second responsibility segment according to the second segment features and the second supplementary features to obtain a second classification result; The first classification result and the second classification result are used together as the classification result.

[0032] For example, each patient may generate multiple candidate responsible segments. To further determine whether the first responsible segment, the second responsible segment, etc., are truly responsible segments, machine learning methods can be used for binary classification. The classifier analyzes whether the lesion possesses the characteristics of a responsible segment, while introducing supplementary features such as brightness information from the cross-sectional image, thereby further improving the accuracy of classification.

[0033] For example, the preset classifier can be a random forest classifier, resulting in the following: Figure 9 The judgment result diagram.

[0034] By combining sagittal and transverse images of the lumbar spine for analysis, the lesion characteristics from different perspectives were fully utilized to accurately determine the location of the responsible segment in multi-segment lumbar spine diseases.

[0035] On the other hand, such as Figure 10 As shown, this application provides a system for determining the responsible segment in lumbar spine diseases, the system comprising: The data acquisition module 201 is used to acquire image data to be identified, wherein the image data to be identified includes a sagittal image of the lumbar spine and a cross-sectional image of the lumbar spine; The first annotation module 202 is used to annotate the sagittal image of the lumbar spine according to the first annotation information of each segment of the lumbar spine and the second annotation information of the lesion on the lumbar spine, so as to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information. The first identification module 203 is used to identify the first responsible segment based on the first sagittal image and the second sagittal image using a first preset YOLO model; The second annotation module 204 is used to annotate the cross-sectional image based on the third annotation information to obtain the target cross-sectional image; The second identification module 205 is used to identify the second responsible segment based on the target cross-sectional image using a second preset YOLO model; The determination module 206 is used to input the first responsible segment and the second responsible segment into a preset classifier to obtain a classification result, and to annotate the image data to be identified based on the classification result, so as to complete the determination of the responsible segment of lumbar spine disease.

[0036] In one possible implementation, such as Figure 11 As shown, this application embodiment provides a terminal device 300, including: a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it achieves: acquiring image data to be identified, wherein the image data to be identified includes a sagittal image of the lumbar spine and a cross-sectional image of the lumbar spine. Based on the first annotation information of each segment of the lumbar vertebra and the second annotation information of the lesion on the lumbar vertebra, the sagittal image of the lumbar vertebra is annotated to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information; Based on the first sagittal image and the second sagittal image, the first responsible segment is identified using a first preset YOLO model; The cross-sectional image is annotated based on the third annotation information to obtain the target cross-sectional image; Based on the target cross-sectional image, the second responsible segment is identified using a second preset YOLO model; The first and second responsible segments are input into a preset classifier to obtain classification results. The image data to be identified is then labeled based on the classification results to complete the determination of the responsible segment for lumbar spine diseases.

[0037] In one possible implementation, such as Figure 12 As shown, this application embodiment provides a computer-readable storage medium 400, on which a computer program 411 is stored. When the computer program 411 is executed by a processor, it performs the following: acquiring image data to be identified, wherein the image data to be identified includes a sagittal image of the lumbar spine and a cross-sectional image of the lumbar spine. Based on the first annotation information of each segment of the lumbar vertebra and the second annotation information of the lesion on the lumbar vertebra, the sagittal image of the lumbar vertebra is annotated to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information; Based on the first sagittal image and the second sagittal image, the first responsible segment is identified using a first preset YOLO model; The cross-sectional image is annotated based on the third annotation information to obtain the target cross-sectional image; Based on the target cross-sectional image, the second responsible segment is identified using a second preset YOLO model; The first and second responsible segments are input into a preset classifier to obtain classification results. The image data to be identified is then labeled based on the classification results to complete the determination of the responsible segment for lumbar spine diseases.

[0038] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0039] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0040] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0041] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0042] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0043] In the embodiments provided in this application, it should be understood that the disclosed apparatus / network devices and methods can be implemented in other ways. For example, the apparatus / network device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0044] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0045] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

[0046] The above description discloses only preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for determining a responsible segment of lumbar spinal disease, characterized by, The method comprises: acquiring to-be-identified image data, wherein the to-be-identified image data comprises a sagittal image of a lumbar vertebra and a cross-sectional image of the lumbar vertebra; annotating the sagittal image of the lumbar vertebra according to first annotation information of each segment of the lumbar vertebra and second annotation information of a lesion on the lumbar vertebra, to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information; identifying a first responsible segment by a first preset YOLO model according to the first sagittal image and the second sagittal image; annotating the cross-sectional image based on the third annotation information to obtain a target cross-sectional image; identifying a second responsible segment by a second preset YOLO model according to the target cross-sectional image; inputting the first responsible segment and the second responsible segment into a preset classifier to obtain a classification result, and annotating the to-be-identified image data based on the classification result to complete the determination of the responsible segment of the lumbar vertebra disease.

2. The method for determining a responsible segment of a lumbar disease according to claim 1, wherein The step of annotating the sagittal image of the lumbar vertebra according to the first annotation information of each segment of the lumbar vertebra and the second annotation information of the lesion on the lumbar vertebra comprises: taking the category, center information and bounding box information of each segment of the lumbar vertebra as the first annotation information, and annotating the sagittal image of the lumbar vertebra according to the first annotation information of each segment of the lumbar vertebra to obtain a first sagittal image with the first annotation information; taking the center position information of the lesion on the lumbar vertebra as the second annotation information, and annotating the sagittal image of the lumbar vertebra according to the second annotation information of the lesion on the lumbar vertebra to obtain a second sagittal image with the second annotation information.

3. The method for determining a responsible segment of a lumbar disease according to claim 1, wherein The step of identifying a first responsible segment by a first preset YOLO model according to the first sagittal image and the second sagittal image comprises: inputting the first sagittal image into a first identification unit in the first preset YOLO model to identify a segment on the lumbar vertebra, and taking the segment on the lumbar vertebra identified by the first identification unit as a candidate segment; inputting the second sagittal image into a second identification unit in the first preset YOLO model to identify a lesion on the lumbar vertebra, and taking the lesion on the lumbar vertebra identified by the second identification unit as a candidate segment; determining a candidate segment having a preset correlation with the target lesion based on the second annotation information corresponding to the target lesion and the first annotation information corresponding to the candidate segment, and taking the candidate segment having the preset correlation with the target lesion as the first responsible segment.

4. The method for determining a responsible segment of lumbar disease according to claim 3, wherein The step of determining a candidate segment having a preset correlation with the target lesion based on the second annotation information corresponding to the target lesion and the first annotation information corresponding to the candidate segment, and taking the candidate segment having the preset correlation with the target lesion as the first responsible segment comprises: The second annotation information corresponding to the target lesion is used as the first discrimination data, and the first annotation information corresponding to the candidate segment is used as the second discrimination data. The first discrimination data and the second discrimination data are matched to obtain the matching result. Based on the matching results, candidate segments that have a preset association with the target lesion are determined, and the candidate segments that have a preset association with the target lesion are designated as the first responsible segments.

5. The method for determining a responsible segment of lumbar disease according to claim 4, wherein The step of matching the first discrimination data and the second discrimination data to determine candidate segments that have a preset association with the target lesion, and selecting the candidate segments that have a preset association with the target lesion as the first responsible segment, includes: When the matching result indicates that the target lesion is within any candidate segment, the candidate segment where the target lesion is located is taken as a candidate segment with a preset association relationship with the target lesion, and the candidate segment with a preset association relationship with the target lesion is taken as the first responsible segment; When the matching result indicates that the target lesion is not among all candidate segments, the candidate segment closest to the target lesion is taken as the candidate segment with a preset association relationship with the target lesion, and the candidate segment with the preset association relationship with the target lesion is taken as the first responsible segment.

6. The method for determining a responsible segment of a lumbar disease according to claim 1, wherein The step of annotating the cross-sectional image based on the third annotation information to obtain the target cross-sectional image includes: Obtain historical annotation information from historical cross-sectional images, and use the information category corresponding to the historical annotation information as the information category of the third annotation information; The cross-sectional images are stitched together to obtain the image to be labeled; The cross-sectional image is annotated based on the third annotation information to obtain the target cross-sectional image.

7. The method for determining a responsible segment of a lumbar disease according to claim 1, wherein The step of inputting the first responsibility segment and the second responsibility segment into a preset classifier to obtain a classification result includes: The first responsible segment is input into the feature extraction unit of the preset classifier to obtain the first segment feature and the first supplementary feature of the first responsible segment. The first supplementary feature includes the brightness feature of the cross-sectional image. The classification unit of the preset classifier classifies the first responsible segment according to the first segment features and the first supplementary features to obtain a first classification result; The second responsibility segment is input into the feature extraction unit of the preset classifier to obtain the second segment feature and the second supplementary feature of the second responsibility segment. The second supplementary feature includes the brightness feature of the cross-sectional image. The classification unit of the preset classifier classifies the second responsibility segment according to the second segment features and the second supplementary features to obtain a second classification result; The first classification result and the second classification result are used together as the classification result.

8. A responsible segment determination system for lumbar disc disease, characterized by, The system includes: The data acquisition module is used to acquire image data to be identified, wherein the image data to be identified includes a sagittal image of the lumbar spine and a cross-sectional image of the lumbar spine; The first annotation module is used to annotate the sagittal image of the lumbar spine according to the first annotation information of each segment of the lumbar spine and the second annotation information of the lesion on the lumbar spine, so as to obtain a first sagittal image with the first annotation information and a second sagittal image with the second annotation information. The first identification module is used to identify the first responsible segment based on the first sagittal image and the second sagittal image using a first preset YOLO model; The second annotation module is used to annotate the cross-sectional image based on the third annotation information to obtain the target cross-sectional image; The second identification module is used to identify the second responsible segment based on the target cross-sectional image using a second preset YOLO model; The determination module is used to input the first responsible segment and the second responsible segment into a preset classifier to obtain a classification result, and to annotate the image data to be identified based on the classification result in order to complete the determination of the responsible segment of lumbar spine disease.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for determining the responsible segment of lumbar spine diseases as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. When the computer program is executed by the processor, it implements the method for determining the responsible segment of lumbar spine diseases as described in any one of claims 1 to 7.

Images