A method for constructing an AI simulation teaching model for temporomandibular joint surgery

Abstract

The present invention relates to the field of image processing technology, and in particular to a method for constructing an AI simulation teaching model for temporomandibular joint surgery, comprising: obtaining key grayscale images of several frames of temporomandibular joint surgery videos according to the difference in pixel grayscale values ​​of grayscale images of temporomandibular joint surgery videos of adjacent frames; dividing the key grayscale image of each frame of temporomandibular joint surgery video into several local areas; dividing all local areas into several marked image blocks and unmarked image blocks; obtaining the visual quality of each marked image block in the key grayscale image of each frame of temporomandibular joint surgery video according to the difference in pixel grayscale values ​​of the marked image blocks and the unmarked image blocks in the key grayscale image of each frame of temporomandibular joint surgery video; performing resolution enhancement on the key grayscale image of the temporomandibular joint surgery video, and constructing an AI simulation teaching model for temporomandibular joint surgery. The present invention improves the accuracy of the AI ​​teaching model for temporomandibular joint surgery.

Description

Technical Field

[0001] The present invention relates to the field of image data processing, and in particular to a method for constructing an AI simulation teaching model for temporomandibular joint surgery. Background Art

[0002] The temporomandibular joint is one of the most complex joints in the human body, responsible for functions such as chewing and speech. Temporomandibular joint surgery involves a variety of surgical types, such as disc repair and joint replacement, which is of great significance for treating joint diseases, reconstructing joint function and improving the quality of life of patients. At present, temporomandibular joint surgery mainly relies on manual operation by experienced surgeons, and the results of the surgery depend largely on the experience and personal skills of the doctors. There is a lack of effective simulation and training tools. Young doctors often need to accumulate experience in actual surgery, which not only increases the surgical risk of patients, but also prolongs the growth cycle of doctors. With the rapid development of artificial intelligence technology, by building an AI simulation teaching model for temporomandibular joint surgery, learners can practice surgical operations without risk and conduct real-time evaluation and feedback on the results of the surgical operations. In the process of building the model, it is necessary to collect clear images from the temporomandibular joint surgery video. Since the image captured from the video will reduce the resolution of the captured image, the clarity of the captured image is not enough, so that the subtle structure and details cannot be accurately captured during the model construction process, which leads to certain errors in the teaching model and cannot effectively improve the safety and success rate of the surgery. Summary of the invention

[0003] The present invention provides a method for constructing an AI simulation teaching model for temporomandibular joint surgery, so as to solve the problem that capturing images from videos will reduce the resolution of the captured images, resulting in insufficient clarity of the captured images, making it impossible to accurately capture subtle structures and details during the model construction process, thereby causing certain errors in constructing the teaching model.

[0004] The present invention provides a method for constructing an AI simulation teaching model for temporomandibular joint surgery using the following technical solutions:

[0005] An embodiment of the present invention provides a method for constructing an AI simulation teaching model for temporomandibular joint surgery, the method comprising the following steps:

[0006] Acquire several frames of temporomandibular joint surgery video grayscale images and the time node corresponding to each frame of temporomandibular joint surgery video grayscale image;

[0007] According to the difference of the grayscale values ​​of the pixels in the grayscale images of the temporomandibular joint surgery videos of adjacent frames, the key grayscale images of several frames of the temporomandibular joint surgery videos are obtained;

[0008] Divide the key grayscale image of each frame of the temporomandibular joint surgery video into several local areas; divide all the local areas into several marked image blocks and unmarked image blocks;

[0009] According to the difference in the grayscale values ​​of pixels in the marked image blocks and the unmarked image blocks in the key grayscale image of each frame of the temporomandibular joint surgery video, the visual quality of each marked image block in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained;

[0010] Segment the temporomandibular joint cartilage area in the key grayscale image of each frame of the temporomandibular joint surgery video, and obtain the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video according to the slope value of the straight line passing through two adjacent pixel points on the contour boundary line of the temporomandibular joint cartilage surface and the visual quality of each marked image block;

[0011] According to the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video and the time node of the grayscale image of each frame of the temporomandibular joint surgery video, the resolution enhancement credibility of the key grayscale image of each frame of the temporomandibular joint surgery video is obtained;

[0012] The credibility is increased according to the resolution of the key grayscale image of each frame of the temporomandibular joint surgery video, the resolution of the key grayscale image of the temporomandibular joint surgery video is enhanced, and an AI simulation teaching model of temporomandibular joint surgery is constructed.

[0013] Preferably, the step of obtaining the key grayscale images of several frames of temporomandibular joint surgery videos according to the difference in the grayscale values ​​of the pixels of the grayscale images of the temporomandibular joint surgery videos in adjacent frames comprises the following specific steps:

[0014] According to the difference of the grayscale values ​​of the pixels in the grayscale images of the temporomandibular joint surgery videos in adjacent frames and the difference of the information entropy of the grayscale values ​​of the pixels, the information difference between the grayscale images of the temporomandibular joint surgery videos in adjacent frames is obtained;

[0015] According to the information difference between the grayscale images of adjacent frames of temporomandibular joint surgery video, the key grayscale images of several frames of temporomandibular joint surgery video are obtained.

[0016] Preferably, the information difference between the grayscale images of the temporomandibular joint surgery videos of adjacent frames is obtained based on the difference in the grayscale values ​​of the pixels in the grayscale images of the temporomandibular joint surgery videos of adjacent frames and the information entropy difference of the grayscale values ​​of the pixels, and the specific formula included is as follows:

[0017]

[0018] In the formula, Indicates Frame and Information difference between frame temporomandibular joint surgery video grayscale images; Represents the number of all pixels in the grayscale image of the temporomandibular joint surgery video; Indicates The grayscale image of the temporomandibular joint surgery video The gray value of each pixel; Indicates The grayscale image of the temporomandibular joint surgery video The gray value of each pixel; Indicates The information entropy of the grayscale values ​​of all pixels in the grayscale image of the frame temporomandibular joint surgery video; Indicates The information entropy of the grayscale values ​​of all pixels in the grayscale image of the frame temporomandibular joint surgery video; is the absolute value function.

[0019] Preferably, the step of obtaining key grayscale images of several frames of temporomandibular joint surgery videos according to the information difference between grayscale images of adjacent frames of temporomandibular joint surgery videos comprises the following specific steps:

[0020] In all frames of temporomandibular joint surgery video grayscale images, the information differences between all two adjacent frames of temporomandibular joint surgery video grayscale images are sorted from small to large to obtain an information difference sequence;

[0021] In the information difference sequence, the absolute value of the difference between two adjacent information differences is calculated, and the two frames of temporomandibular joint surgery video grayscale images corresponding to the previous information difference of the two information differences corresponding to the maximum value of the absolute value of the difference between all adjacent information differences are recorded as the key grayscale images of the temporomandibular joint surgery video.

[0022] Preferably, the step of dividing all local areas into a plurality of marked image blocks and unmarked image blocks includes the following specific steps:

[0023] Acquire shape template images of several surgical instruments;

[0024] Using template matching technology, the shape template images of all surgical instruments are matched with all local areas of the key grayscale image of the temporomandibular joint surgery video. The local area where the surgical instrument exists in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained by matching, and recorded as the target local area of ​​the surgical instrument;

[0025] From the target local area of ​​the surgical instrument, the target local area of ​​the surgical instrument including the temporomandibular joint cartilage area is screened out and recorded as a marked image block; the target local area of ​​the surgical instrument not including the temporomandibular joint area is recorded as an unlabeled image block.

[0026] Preferably, the visual quality of each marked image block in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained according to the difference in the grayscale values ​​of the pixels in the marked image block and the unmarked image block in the key grayscale image of each frame of the temporomandibular joint surgery video, and the specific steps include the following:

[0027]

[0028] In the formula, Indicates The key grayscale image of the temporomandibular joint surgery video The visual quality of the labeled image patches; Indicates The number of all unlabeled image patches in the key grayscale image of the frame TMJ surgery video; Indicates The key grayscale image of the temporomandibular joint surgery video The average gray value of all pixels in the marked image block; Indicates The key grayscale image of the temporomandibular joint surgery video The average gray value of all pixels in an unlabeled image block; is the absolute value function.

[0029] Preferably, the step of obtaining the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video according to the slope value of the straight line passing through two adjacent pixel points on the contour boundary line of the temporomandibular joint cartilage surface and the visual quality of each marked image block comprises the following specific steps:

[0030] Perform edge detection on the key grayscale image of each frame of the temporomandibular joint surgery video to obtain a number of edge pixel points;

[0031] According to the slope value of the straight line passing through two adjacent pixel points on the contour edge line of the temporomandibular joint cartilage area and the edge pixel points, a slope sequence and the number of breakpoints are obtained;

[0032] According to the number of breakpoints on the contour boundary line of the temporomandibular joint cartilage surface, the slope sequence and the visual quality of each marked image block, the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained.

[0033] Preferably, the method of obtaining a slope sequence and the number of breakpoints according to the slope value of a straight line passing through two adjacent pixel points on the contour edge line of the temporomandibular joint cartilage region and the edge pixel points comprises the following specific steps:

[0034] The number of pixels on the contour boundary of the temporomandibular joint cartilage area that are not edge pixels is recorded as the number of breakpoints;

[0035] In the key grayscale image of any frame of temporomandibular joint surgery video, starting from any pixel point on the contour edge line of the temporomandibular joint cartilage area, the slopes of the straight lines passing through two adjacent pixel points are counted in the clockwise direction to obtain a slope sequence.

[0036] Preferably, the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained according to the number of breakpoints on the contour boundary line of the temporomandibular joint cartilage surface, the slope sequence and the visual quality of each marked image block, including the specific formula as follows:

[0037]

[0038] In the formula, Indicates The contrast enhancement requirement of local areas in the key grayscale images of the frame-by-frame temporomandibular joint surgery video; Indicates The key grayscale image of the temporomandibular joint surgery video The visual quality of the labeled image patches; Indicates The number of all labeled image patches in the key grayscale image of the frame TMJ surgery video; Indicates The number of breakpoints on the contour boundary line of the temporomandibular joint cartilage area in the key grayscale image of the frame temporomandibular joint surgery video; Indicates The slope sequence corresponding to the key grayscale image of the temporomandibular joint surgery video Slope value; Indicates The slope sequence corresponding to the key grayscale image of the temporomandibular joint surgery video Slope value; Indicates The number of all slope values ​​on the slope sequence corresponding to the key grayscale image of the frame temporomandibular joint surgery video; is the absolute value function.

[0039] Preferably, the resolution enhancement credibility of the key grayscale image of each frame of the temporomandibular joint surgery video is obtained according to the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video and the time node of the grayscale image of each frame of the temporomandibular joint surgery video, and the specific formula included is as follows:

[0040]

[0041] In the formula, Indicates The resolution of the key grayscale images of the frame temporomandibular joint surgery video increases the credibility; Indicates The contrast enhancement requirement of local areas in the key grayscale images of the frame-by-frame temporomandibular joint surgery video; Indicates The time nodes of the key grayscale images of the frame temporomandibular joint surgery video; Indicates the time node of the first frame of the temporomandibular joint surgery video grayscale image; Indicates the time node of the last frame of the temporomandibular joint surgery video grayscale image; is the hyperbolic tangent function.

[0042] The beneficial effects of the technical solution of the present invention are: according to the difference in grayscale values ​​of pixels in grayscale images of adjacent frames of temporomandibular joint surgery videos, key grayscale images of several frames of temporomandibular joint surgery videos are obtained; the key grayscale image of each frame of temporomandibular joint surgery video is equally divided into several local areas; all local areas are divided into several marked image blocks and unmarked image blocks; according to the difference in grayscale values ​​of pixels in marked image blocks and unmarked image blocks in the key grayscale image of each frame of temporomandibular joint surgery video, the visual quality of each marked image block in the key grayscale image of each frame of temporomandibular joint surgery video is obtained; better data support is provided for the construction of an AI teaching model for temporomandibular joint surgery, so that the teaching model can restore the detailed changes in the real surgical process to the greatest extent during application. According to the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video and the time node of the grayscale image of each frame of the temporomandibular joint surgery video, the resolution enhancement credibility of the key grayscale image of each frame of the temporomandibular joint surgery video is obtained; according to the resolution enhancement credibility of the key grayscale image of each frame of the temporomandibular joint surgery video, the key grayscale image of the temporomandibular joint surgery video is enhanced, and an AI simulation teaching model for temporomandibular joint surgery is constructed. The accuracy of the AI ​​simulation teaching model for temporomandibular joint surgery is improved, providing an efficient, safe and reliable surgery learning platform for medical students and doctors. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1This is a flowchart of the steps of a method for constructing an AI simulation teaching model for temporomandibular joint surgery according to the present invention. DETAILED DESCRIPTION

[0045] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following is a method for constructing an AI simulation teaching model for temporomandibular joint surgery proposed by the present invention, in combination with the accompanying drawings and preferred embodiments, and its specific implementation method, structure, characteristics and effects are described in detail as follows. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0046] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0047] The specific scheme of the method for constructing an AI simulation teaching model for temporomandibular joint surgery provided by the present invention is described in detail below with reference to the accompanying drawings.

[0048] See also Figure 1 , which shows a flowchart of a method for constructing an AI simulation teaching model for temporomandibular joint surgery provided by an embodiment of the present invention, the method comprising the following steps:

[0049] Step S001: Acquire several frames of temporomandibular joint surgery video grayscale images and shape template images of surgical instruments as well as the time nodes corresponding to each frame of temporomandibular joint surgery video grayscale images.

[0050] It should be noted that the teaching model is a complex and systematic process. First, the goal and scope of the teaching model are determined, including the target audience, teaching content, and teaching effectiveness evaluation criteria. Then, relevant data on temporomandibular joint surgery are collected, including surgical videos, anatomical model data, medical literature, etc. Then, the structure and function of the teaching model are designed. According to the designed model structure, the model is developed using appropriate technologies and tools. Finally, the developed model is integrated into the teaching platform or equipment and deployed.

[0051] By using the public temporomandibular joint surgery video learning materials, each frame image of the temporomandibular joint surgery video is processed in grayscale to obtain several frames of temporomandibular joint surgery video grayscale images and the time node corresponding to each frame of the temporomandibular joint surgery video grayscale image.

[0052] A plurality of frames of shape template images of surgical instruments are obtained from a medical database.

[0053] What needs to be explained is that when obtaining surgical video images, attention must be paid to protecting patient privacy and data security, ensuring that the process of obtaining video images is compliant and legal and does not infringe on the patient's privacy rights.

[0054] Step S002: obtaining key grayscale images of several frames of temporomandibular joint surgery video according to the differences in the grayscale values ​​of the pixels of the grayscale images of adjacent frames of the temporomandibular joint surgery video.

[0055] It should be noted that in the method of constructing an AI simulation teaching model for temporomandibular joint surgery, by enhancing the contrast and brightness of the video frame images, the model can more easily distinguish key details in the surgical process, such as surgical tools, tissue structures, and boundaries of the surgical area; through image enhancement technology, diverse training data can be generated, the generalization ability of the model can be improved, and the learner's visual experience can be effectively improved, making it easier for them to understand the complexity and details of the surgical process.

[0056] It should be noted that, since temporomandibular joint surgery includes many types of surgeries, each type of surgery corresponds to the corresponding temporomandibular joint pathology; based on the numerous surgical videos in the resource library, surgical videos of one type of surgery are selected as the analysis object, and the videos of this type of surgery are divided into multiple image data in frames.

[0057] The information entropy of the grayscale value of the pixel is obtained by a formula for calculating the information entropy. The formula for the information entropy is a well-known technology, and the specific method will not be introduced here.

[0058] It should be noted that key images are extracted from the grayscale images of the temporomandibular joint surgery video, and these key images represent the main operations and change points in a short period of time.

[0059] No. Frame and The information difference between the grayscale images of the temporomandibular joint surgery video frames is calculated as follows:

[0060]

[0061] In the formula, Indicates Frame and Information difference between frame temporomandibular joint surgery video grayscale images; Represents the number of all pixels in the grayscale image of the temporomandibular joint surgery video; Indicates The grayscale image of the temporomandibular joint surgery video The gray value of each pixel; Indicates The grayscale image of the temporomandibular joint surgery video The gray value of each pixel; Indicates The information entropy of the grayscale values ​​of all pixels in the grayscale image of the frame temporomandibular joint surgery video; Indicates The information entropy of the grayscale values ​​of all pixels in the grayscale image of the frame temporomandibular joint surgery video; is the absolute value function.

[0062] What needs to be explained is that It represents the average value of the grayscale difference between all pixels at the same position in two adjacent frames of temporomandibular joint surgery video grayscale images. The larger the average value, the greater the difference between the pixels at the same position in the two frames of grayscale images, that is, the greater the difference in change between the two adjacent frames of grayscale images. It indicates the difference of the information entropy of the grayscale value of pixels between two adjacent grayscale images. The larger the difference, the greater the difference in the distribution of pixel types between the two adjacent grayscale images. It represents the information difference between two adjacent grayscale image frames. The larger the product, the greater the information difference between the two adjacent grayscale image frames. That is, in the surgical video, the key steps of the surgical operation or significant changes in the anatomical structure will increase the information difference between the adjacent grayscale image frames.

[0063] So far, the information difference between all two adjacent frames of temporomandibular joint surgery video grayscale images has been obtained.

[0064] In all frames of temporomandibular joint surgery video grayscale images, the information differences between all two adjacent frames of temporomandibular joint surgery video grayscale images are sorted from small to large to obtain an information difference sequence.

[0065] In the information difference sequence, the absolute value of the difference between two adjacent information differences is calculated, and the two frames of temporomandibular joint surgery video grayscale images corresponding to the previous information difference of the two information differences corresponding to the maximum value of the absolute value of the difference between all adjacent information differences are recorded as the key grayscale images of the temporomandibular joint surgery video.

[0066] It should be noted that: if there are multiple maximum values, the first maximum value in the information difference sequence is taken.

[0067] At this point, several frames of key grayscale images of the temporomandibular joint surgery video are obtained.

[0068] Step S003: Divide the key grayscale image of each frame of the temporomandibular joint surgery video into several local areas; divide all the local areas into several marked image blocks and unmarked image blocks; obtain the visual quality of each marked image block in the key grayscale image of each frame of the temporomandibular joint surgery video according to the difference in the grayscale values ​​of the pixels in the marked image blocks and the unmarked image blocks in the key grayscale image of each frame of the temporomandibular joint surgery video; segment the temporomandibular joint cartilage area in the key grayscale image of each frame of the temporomandibular joint surgery video, according to the slope value of the straight line passing through two adjacent pixels on the contour boundary line of the temporomandibular joint cartilage surface and the visual quality of each marked image block; obtain the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video.

[0069] It should be noted that contrast enhancement of key grayscale images in temporomandibular joint surgery videos can more clearly display surgical details and structures, highlight the contrast of local details and regions, and is more helpful for details such as instruments, tissue edges, and blood vessels in surgical videos.

[0070] Edge detection is performed on the key grayscale image of each frame of the temporomandibular joint surgery video to obtain a number of edge pixels. The Canny edge detection algorithm is used for edge detection, which is a well-known technology.

[0071] The embodiment of the present invention adopts a segmentation neural network to identify the temporomandibular joint cartilage area in the key grayscale image of each frame of the temporomandibular joint surgery video.

[0072] The relevant content of the segmentation neural network is as follows:

[0073] The segmentation neural network used in this embodiment is the Mask R-CNN neural network; the data set used is the key grayscale image data set of all frames of temporomandibular joint surgery videos. Mask R-CNN is a well-known technology, and the specific method is not introduced here. The full name of Mask R-CNN in Chinese is "Mask Region-based Convolutional Neural Network", and the full name in English is "Mask Region-based Convolutional Neural Network".

[0074] The pixels that need to be segmented are divided into two categories, that is, the labeling process of the training set is: single-channel semantic label, the corresponding position pixel belongs to the normal area is labeled as 0, and the corresponding position pixel belongs to the temporomandibular joint cartilage area is labeled as 1.

[0075] The task of the network is classification, so the loss function used is the cross entropy loss function.

[0076] The temporomandibular joint cartilage area in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained by segmenting the neural network. This process is a well-known technology and the specific method will not be introduced here.

[0077] The number of pixels on the contour boundary line of the temporomandibular joint cartilage area that are not edge pixels is recorded as the number of breakpoints.

[0078] In the key grayscale image of any frame of temporomandibular joint surgery video, starting from any pixel point on the contour edge line of the temporomandibular joint cartilage area, the slopes of the straight lines passing through two adjacent pixel points are counted in the clockwise direction to obtain a slope sequence.

[0079] The first The key grayscale image of the temporomandibular joint surgery video is equally divided into a number of local areas, where the local area is a square, and the number of the local areas is a preset number threshold of 20, which is used as an example for description;

[0080] Using template matching technology, the shape template images of all surgical instruments are matched with all local areas of the key grayscale image of the temporomandibular joint surgery video. The local area where the surgical instrument exists in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained by matching, and recorded as the target local area of ​​the surgical instrument;

[0081] From the target local area of ​​the surgical instrument, the target local area of ​​the surgical instrument including the temporomandibular joint cartilage area is screened out and recorded as a marked image block; the target local area of ​​the surgical instrument not including the temporomandibular joint area is recorded as an unlabeled image block.

[0082] Among them, the template matching technology is a well-known technology, and the specific method will not be introduced here.

[0083] No. The key grayscale image of the temporomandibular joint surgery video The visual quality of a labeled image patch is calculated as follows:

[0084]

[0085] In the formula, Indicates The key grayscale image of the temporomandibular joint surgery video Visual quality of labeled image patches Indicates The number of all unlabeled image patches in the key grayscale image of the frame TMJ surgery video; Indicates The key grayscale image of the temporomandibular joint surgery video The average gray value of all pixels in the marked image block; Indicates The key grayscale image of the temporomandibular joint surgery video The average gray value of all pixels in an unlabeled image block; is the absolute value function.

[0086] What needs to be explained is that Indicates The number of all unlabeled image patches in the key grayscale image of the frame TMJ surgery video; Indicates The key grayscale image of the temporomandibular joint surgery video The labeled image patch and The larger the value of the sum of the grayscale differences of the pixels between the unmarked image blocks, the greater the grayscale difference between the marked image block and the unmarked image blocks in the frame image, indicating that the brightness of the marked image block in the frame image is more prominent, that is, the visual quality is higher.

[0087] No. The calculation method of the contrast enhancement requirement of the local area in the key grayscale image of the frame temporomandibular joint surgery video is as follows:

[0088]

[0089] In the formula, Indicates The contrast enhancement requirement of local areas in the key grayscale images of the frame-by-frame temporomandibular joint surgery video; Indicates The key grayscale image of the temporomandibular joint surgery video The visual quality of the labeled image patches; Indicates The number of all labeled image patches in the key grayscale image of the frame TMJ surgery video; Indicates The number of breakpoints on the contour boundary line of the temporomandibular joint cartilage area in the key grayscale image of the frame temporomandibular joint surgery video; Indicates The slope sequence corresponding to the key grayscale image of the temporomandibular joint surgery video Slope value; Indicates The slope sequence corresponding to the key grayscale image of the temporomandibular joint surgery video Slope value; Indicates The number of all slope values ​​on the slope sequence corresponding to the key grayscale image of the frame temporomandibular joint surgery video; is the absolute value function.

[0090] What needs to be explained is that when the denominator in the formula is 0, set the denominator to 1 and use this as an example to ensure that the formula is valid.

[0091] What needs to be explained is that Indicates the smoothness of the contour edge line of the mandibular joint cartilage surface in the key grayscale image of the temporomandibular joint surgery video. The larger the average value, the more uneven the contour edge line is. In low-contrast images, the contour edge line of the cartilage is not clear enough in the image, resulting in unevenness. It indicates the number of breakpoints on the edge line. The more breakpoints there are, the more discontinuous the edge line is. The discontinuity of the contour edge line of the mandibular articular cartilage surface in the key grayscale image usually indicates that the image contrast is low. It represents the average visual quality index of all the marked image blocks in the grayscale image of this frame. The smaller the average value is, the smaller the difference in information change between the marked image blocks and the unmarked image blocks is, that is, the lower the visual quality is; It indicates the contrast enhancement requirement of the local area in the key grayscale image. The larger the product is, the lower the contrast of the local area in the key image is, so a higher contrast is needed to highlight the contrast of the key image.

[0092] At this point, the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained.

[0093] Step S004: Based on the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video and the time node of the grayscale image of each frame of the temporomandibular joint surgery video, the resolution increase credibility of the key grayscale image of each frame of the temporomandibular joint surgery video is obtained.

[0094] First, the time nodes of the grayscale images of all frames of the temporomandibular joint surgery video are obtained, and the time node of the first frame of the grayscale image is recorded as ; The time node of the last frame of grayscale image is recorded as ;

[0095] The key grayscale images of the temporomandibular joint surgery video obtained above are recorded at the time node of each frame of the key grayscale image;

[0096] No. The calculation method of the resolution increase credibility of the key grayscale image of the temporomandibular joint surgery video frame is as follows:

[0097]

[0098] In the formula, Indicates The resolution of the key grayscale images of the frame temporomandibular joint surgery video increases the credibility; Indicates The contrast enhancement requirement of local areas in the key grayscale images of the frame-by-frame temporomandibular joint surgery video; Indicates The time nodes of the key grayscale images of the frame temporomandibular joint surgery video; Indicates the time node of the first frame of the temporomandibular joint surgery video grayscale image; Indicates the time node of the last frame of the temporomandibular joint surgery video grayscale image; is the hyperbolic tangent function.

[0099] What needs to be explained is that Indicates The difference between the time node of the key grayscale image of the frame temporomandibular joint surgery video and the time node of the grayscale image of the first frame temporomandibular joint surgery video. The time node representing the grayscale image of the last frame of temporomandibular joint surgery video and the The difference between the time nodes of the key grayscale images of the temporomandibular joint surgery video. The adaptive interpolation weight of the key grayscale image of the temporomandibular joint surgery video is represented. The larger the product, the farther the time node of the key grayscale image is from the time node of the start of the surgery and the time node of the end of the surgery, which means that the time node of the key grayscale image is in the surgery. The details of the surgery at this time node are relatively important, so a higher difference weight is required; It indicates the credibility of the resolution increase of the key grayscale image. The larger the product is, the higher the credibility of the resolution increase of the key grayscale image is, that is, the resolution of the key grayscale image can better highlight the detail changes during the operation.

[0100] At this point, the resolution of the key grayscale image of each frame of the temporomandibular joint surgery video is increased with confidence.

[0101] Step S005: Increase the credibility according to the resolution of the key grayscale image of each frame of the temporomandibular joint surgery video, enhance the resolution of the key grayscale image of the temporomandibular joint surgery video, and construct an AI simulation teaching model for temporomandibular joint surgery.

[0102] According to the resolution increase of the key grayscale image of each frame of the temporomandibular joint surgery video, the image pyramid algorithm was used to enhance the resolution of the key grayscale image of the temporomandibular joint surgery video to obtain the enhanced image; and then the finite element simulation method was used to construct the AI ​​simulation teaching model of temporomandibular joint surgery.

[0103] Among them, the image pyramid algorithm and the finite element simulation method are well-known technologies, and the specific methods are not introduced here.

[0104] What needs to be explained is that, for the key grayscale images of all frames of temporomandibular joint surgery videos, blank cells are added to adjacent rows and columns in the image matrix, and the size of the images is ensured to be consistent. The Gaussian kernel is used as the interpolation kernel in the adaptive interpolation process, and the resolution increase credibility of the key grayscale image of each frame of temporomandibular joint surgery video is used as the weight of the interpolation kernel to complete the adaptive interpolation to enhance the resolution of the key image of the temporomandibular joint surgery video and obtain a high-resolution grayscale image of the temporomandibular joint surgery video. Combined with the doctor's surgical movements and steps, the model is established using simulation technology and integrated into interactive learning using AR technology.

[0105] What needs to be explained is that the Gaussian kernel will generate a set of weights of the convolution kernel that obey the standard deviation, and then use them as the weight values ​​for convolving the blank grids in the key grayscale image of the temporomandibular joint surgery video with the neighboring pixels, thereby obtaining the pixel values ​​of the pixels represented by the inserted blank grids. The adaptive interpolation of the key grayscale image is completed to enhance the resolution of the key image. Among them, the convolution kernel of the Gaussian filter of the Gaussian filter is the Gaussian kernel. The Gaussian kernel is a well-known technology, and the specific method will not be introduced here.

[0106] At this point, this embodiment is completed.

[0107] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing an AI simulation teaching model for temporomandibular joint surgery, characterized in that: The method comprises the following steps: Acquire several frames of temporomandibular joint surgery video grayscale images and the time node corresponding to each frame of temporomandibular joint surgery video grayscale image; According to the difference of the grayscale values ​​of the pixels in the grayscale images of the temporomandibular joint surgery videos of adjacent frames, the key grayscale images of several frames of the temporomandibular joint surgery videos are obtained; Divide the key grayscale image of each frame of the temporomandibular joint surgery video into several local areas; divide all the local areas into several marked image blocks and unmarked image blocks; According to the difference in the grayscale values ​​of pixels in the marked image blocks and the unmarked image blocks in the key grayscale image of each frame of the temporomandibular joint surgery video, the visual quality of each marked image block in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained; Segment the temporomandibular joint cartilage area in the key grayscale image of each frame of the temporomandibular joint surgery video, and obtain the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video according to the slope value of the straight line passing through two adjacent pixel points on the contour boundary line of the temporomandibular joint cartilage surface and the visual quality of each marked image block; According to the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video and the time node of the grayscale image of each frame of the temporomandibular joint surgery video, the resolution enhancement credibility of the key grayscale image of each frame of the temporomandibular joint surgery video is obtained; The credibility of the key grayscale image in each frame of the temporomandibular joint surgery video is increased by resolution enhancement, and the key grayscale image in the temporomandibular joint surgery video is enhanced, and an AI simulation teaching model for temporomandibular joint surgery is constructed; The specific steps of dividing all local areas into a number of marked image blocks and unmarked image blocks are as follows: Acquire shape template images of several surgical instruments; Using template matching technology, the shape template images of all surgical instruments are matched with all local areas of the key grayscale image of the temporomandibular joint surgery video. The local area where the surgical instrument exists in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained by matching, and recorded as the target local area of ​​the surgical instrument; From the target local area of ​​the surgical instrument, the target local area of ​​the surgical instrument including the temporomandibular joint cartilage area is screened out and recorded as a marked image block; the target local area of ​​the surgical instrument not including the temporomandibular joint area is recorded as an unmarked image block; The method of obtaining the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video according to the slope value of the straight line passing through two adjacent pixel points on the contour boundary line of the temporomandibular joint cartilage surface and the visual quality of each marked image block includes the following specific steps: Perform edge detection on the key grayscale image of each frame of the temporomandibular joint surgery video to obtain a number of edge pixel points; According to the slope value of the straight line passing through two adjacent pixel points on the contour edge line of the temporomandibular joint cartilage area and the edge pixel points, a slope sequence and the number of breakpoints are obtained; According to the number of breakpoints on the contour boundary line of the temporomandibular joint cartilage surface, the slope sequence and the visual quality of each marked image block, the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained; The specific steps of obtaining a slope sequence and a breakpoint number according to the slope value of a straight line passing through two adjacent pixel points on the contour edge line of the temporomandibular joint cartilage region and the edge pixel points are as follows: The number of pixels on the contour boundary of the temporomandibular joint cartilage area that are not edge pixels is recorded as the number of breakpoints; In any frame of the key grayscale image of the temporomandibular joint surgery video, starting from any pixel point on the contour edge line of the temporomandibular joint cartilage area, the slopes of the straight lines passing through two adjacent pixel points are counted in the clockwise direction to obtain a slope sequence; According to the number of breakpoints on the contour boundary line of the temporomandibular joint cartilage surface, the slope sequence and the visual quality of each marked image block, the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained, and the specific formula included is as follows: In the formula, Indicates The contrast enhancement requirement of local areas in the key grayscale images of the frame-by-frame temporomandibular joint surgery video; Indicates The key grayscale image of the temporomandibular joint surgery video The visual quality of the labeled image patches; Indicates The number of all labeled image patches in the key grayscale image of the frame TMJ surgery video; Indicates The number of breakpoints on the contour boundary line of the temporomandibular joint cartilage area in the key grayscale image of the frame temporomandibular joint surgery video; Indicates The slope sequence corresponding to the key grayscale image of the temporomandibular joint surgery video Slope value; Indicates The slope sequence corresponding to the key grayscale image of the temporomandibular joint surgery video Slope value; Indicates The number of all slope values ​​on the slope sequence corresponding to the key grayscale image of the frame temporomandibular joint surgery video; is the absolute value function; The resolution enhancement credibility of the key grayscale image of each frame of the temporomandibular joint surgery video is obtained according to the contrast enhancement requirement of the local area in the key grayscale image of each frame of the temporomandibular joint surgery video and the time node of the grayscale image of each frame of the temporomandibular joint surgery video, including the specific formula as follows: In the formula, Indicates The resolution of the key grayscale images of the frame temporomandibular joint surgery video increases the credibility; Indicates The contrast enhancement requirement of local areas in the key grayscale images of the frame-by-frame temporomandibular joint surgery video; Indicates The time nodes of the key grayscale images of the frame temporomandibular joint surgery video; Indicates the time node of the first frame of the temporomandibular joint surgery video grayscale image; Indicates the time node of the last frame of the temporomandibular joint surgery video grayscale image; is the hyperbolic tangent function.

2. According to claim 1, a method for constructing an AI simulation teaching model for temporomandibular joint surgery is characterized in that: The method of obtaining the key grayscale images of several frames of temporomandibular joint surgery videos according to the difference in the grayscale values ​​of the pixels of the grayscale images of the temporomandibular joint surgery videos in adjacent frames comprises the following specific steps: According to the difference of the grayscale values ​​of the pixels in the grayscale images of the temporomandibular joint surgery videos in adjacent frames and the difference of the information entropy of the grayscale values ​​of the pixels, the information difference between the grayscale images of the temporomandibular joint surgery videos in adjacent frames is obtained; According to the information difference between the grayscale images of adjacent frames of temporomandibular joint surgery video, the key grayscale images of several frames of temporomandibular joint surgery video are obtained.

3. According to claim 2, a method for constructing an AI simulation teaching model for temporomandibular joint surgery is characterized in that: The information difference between the grayscale images of the temporomandibular joint surgery videos of adjacent frames is obtained according to the difference in the grayscale values ​​of the pixels of the temporomandibular joint surgery videos of adjacent frames and the information entropy difference of the grayscale values ​​of the pixels, and the specific formula included is as follows: In the formula, Indicates Frame and Information difference between frame temporomandibular joint surgery video grayscale images; Represents the number of all pixels in the grayscale image of the temporomandibular joint surgery video; Indicates The grayscale image of the temporomandibular joint surgery video The gray value of each pixel; Indicates The grayscale image of the temporomandibular joint surgery video The gray value of each pixel; Indicates The information entropy of the grayscale values ​​of all pixels in the grayscale image of the frame temporomandibular joint surgery video; Indicates The information entropy of the grayscale values ​​of all pixels in the grayscale image of the frame temporomandibular joint surgery video; is the absolute value function.

4. According to claim 2, a method for constructing an AI simulation teaching model for temporomandibular joint surgery is characterized in that: The step of obtaining key grayscale images of several frames of temporomandibular joint surgery videos according to the information difference between grayscale images of adjacent frames of temporomandibular joint surgery videos includes the following specific steps: In all frames of temporomandibular joint surgery video grayscale images, the information differences between all two adjacent frames of temporomandibular joint surgery video grayscale images are sorted from small to large to obtain an information difference sequence; In the information difference sequence, the absolute value of the difference between two adjacent information differences is calculated, and the two frames of temporomandibular joint surgery video grayscale images corresponding to the previous information difference of the two information differences corresponding to the maximum value of the absolute value of the difference between all adjacent information differences are recorded as the key grayscale images of the temporomandibular joint surgery video.

5. According to claim 1, a method for constructing an AI simulation teaching model for temporomandibular joint surgery is characterized in that: The visual quality of each marked image block in the key grayscale image of each frame of the temporomandibular joint surgery video is obtained according to the difference in the grayscale value of the pixel points in the marked image block and the unmarked image block in the key grayscale image of each frame of the temporomandibular joint surgery video, and the specific steps include the following: In the formula, Indicates The key grayscale image of the temporomandibular joint surgery video The visual quality of the labeled image patches; Indicates The number of all unlabeled image patches in the key grayscale image of the frame TMJ surgery video; Indicates The key grayscale image of the temporomandibular joint surgery video The average gray value of all pixels in the marked image block; Indicates The key grayscale image of the temporomandibular joint surgery video The average gray value of all pixels in an unlabeled image block; is the absolute value function.

Images