Method and system for identifying the type of impacted wisdom teeth

Abstract

The present invention provides a method and system for identifying the type of impacted wisdom teeth, belonging to the field of image processing technology, obtaining an oral periapical film image to be identified; using a trained wisdom tooth segmentation model to process the acquired oral periapical film image to be identified to obtain a segmented image; performing wisdom tooth region cropping on the segmented image to obtain a wisdom tooth region image; using a trained impaction type identification model to process the wisdom tooth region image to obtain a wisdom tooth impaction type identification result. The present invention uses a two-stage deep learning method to segment wisdom teeth and identify impaction types respectively, and adopts multi-channel input, which can quickly and accurately identify the impaction type of wisdom teeth in oral periapical film images, and ensures classification accuracy.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method and system for identifying the type of impacted wisdom teeth based on oral periapical radiographs based on deep learning. Background Art

[0002] Wisdom teeth usually erupt after the age of eighteen, which is when they become adults. This is why they are called wisdom teeth. Some people's wisdom teeth cannot erupt completely or not at all due to insufficient bone mass in the jaw, which forms impacted teeth. Impacted wisdom teeth are called impacted wisdom teeth.

[0003] Impacted wisdom teeth are the most common impacted teeth. They are very harmful to human health, such as recurrent pericoronitis, dental caries, jaw pain, bruxism, trigeminal neuralgia, malocclusion, absorption of adjacent teeth and supporting tissues, craniomandibular disorders, cystic lesions around wisdom teeth, precancerous lesions of oral mucosa, etc. In order to reduce the harm caused by impacted wisdom teeth to human physical and mental health, they need to be properly treated. Some impacted wisdom teeth need to be removed, while others do not.

[0004] Accurate judgment of the type of impacted wisdom teeth can provide important diagnostic basis for doctors on how to properly deal with impacted wisdom teeth.

[0005] The Chinese invention patent application with application number 202110442986.8 discloses a method for segmenting wisdom teeth in oral curved segment images based on YOLO and U-Net. The oral curved segment images are preprocessed, the positions of the wisdom teeth in the preprocessed images are marked to obtain position labels, and the training set and test set are divided. The YOLO model is trained using the images and position labels, and the confidence threshold of the spatial position is set to obtain the spatial position information of the wisdom teeth; slice processing is performed based on the spatial position information of the wisdom teeth, and all the slices containing the wisdom teeth are preprocessed. After preprocessing, the slices of the training set are annotated at the pixel level; finally, the slice images and position labels of the training set are used to train the U-Net model, and the confidence threshold of the pixel category is set for binarization, and finally the pixel-level classification information of the wisdom teeth is obtained to complete the segmentation of the wisdom teeth.

[0006] Although the above patent application discloses the trained YOLO model and U-Net model, firstly the spatial position information of the wisdom tooth is obtained by using the YOLO model, and then the binary image containing only the foreground and background is obtained by using the U-Net model, and finally the accurate segmentation of the wisdom tooth is achieved. However, the patent application does not realize how to use the segmented wisdom tooth image to classify and identify the impaction type. Summary of the invention

[0007] The purpose of the present invention is to provide a method and system for identifying the type of impacted wisdom teeth based on deep learning, which utilizes a two-stage deep learning method to segment wisdom teeth and identify the impacted type, adopts multi-channel input, and improves the classification accuracy, so as to solve at least one technical problem existing in the above-mentioned background technology.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] In one aspect, the present invention provides a method for identifying the type of impacted wisdom teeth, comprising:

[0010] Acquire an oral periapical radiograph image to be identified;

[0011] The trained wisdom tooth segmentation model is used to process the acquired oral periapical radiograph image to be identified to obtain a segmented image;

[0012] Crop the wisdom tooth region of the segmented image to obtain a wisdom tooth region image;

[0013] The trained impaction type recognition model is used to process the wisdom tooth area image to obtain the wisdom tooth impaction type recognition result.

[0014] Preferably, training the wisdom tooth segmentation model includes: performing feature enhancement preprocessing on the tooth area in multiple oral periapical film images; marking the wisdom teeth and the types of wisdom tooth impaction in the preprocessed images; normalizing the size of the labeled image data, dividing it into a training set, a validation set, and a test set, training the basic network and optimizing the parameters to obtain the final trained wisdom tooth segmentation model.

[0015] Preferably, training the impaction type recognition model includes: based on the segmentation result of the wisdom tooth segmentation model, cropping the wisdom tooth area of ​​the data set image to obtain the wisdom tooth area image; normalizing the size of the cropped wisdom tooth area image data set with the wisdom tooth impaction type annotation, dividing it into a training set, a validation set and a test set, training the basic network and optimizing the parameters to obtain the final trained impaction type recognition model.

[0016] Preferably, gray value normalization is used to perform feature enhancement preprocessing on the tooth area in the oral periapical radiograph image.

[0017] Preferably, the wisdom tooth impaction type labels include mesial impaction, distal impaction, horizontal impaction, vertical impaction and inverted impaction.

[0018] Preferably, the basic network used to train the wisdom tooth segmentation model is:

[0019] One of the semantic segmentation networks FCN, DeepLab or U-Net++;

[0020] or,

[0021] One of the object detection networks YOLO, SSD, Fast R-CNN or Faster R-CNN.

[0022] Preferably, the basic network for training the impaction type recognition model is one of VGG, ResNet, ResNeXt, DenseNet, EfficientNet or RegNet.

[0023] In a second aspect, the present invention provides a system for identifying the type of impacted wisdom teeth, comprising:

[0024] An acquisition module, used for acquiring an oral periapical radiograph image to be identified;

[0025] A segmentation module is used to process the acquired oral periapical radiograph image to be identified using the trained wisdom tooth segmentation model to obtain a segmented image;

[0026] A cropping module is used to crop the wisdom tooth region of the segmented image to obtain a wisdom tooth region image;

[0027] The recognition module is used to process the wisdom tooth area image using the trained impaction type recognition model to obtain the wisdom tooth impaction type recognition result.

[0028] In a third aspect, the present invention provides a non-transitory computer-readable storage medium, which is used to store computer instructions. When the computer instructions are executed by a processor, the method for identifying the type of impacted wisdom teeth as described above is implemented.

[0029] In a fourth aspect, the present invention provides an electronic device, comprising: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes instructions for implementing the method for identifying the type of impacted wisdom teeth as described above.

[0030] The beneficial effects of the present invention are as follows: a two-stage deep learning method is used to segment wisdom teeth and identify impacted types respectively, and multi-channel input is adopted, which can quickly and accurately identify the impacted types of wisdom teeth in oral periapical images, thereby ensuring classification accuracy.

[0031] Additional aspects and advantages of the present invention will be given in part in the following description, which will become obvious from the following description, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0033] Figure 1 This is a functional principle block diagram of the wisdom tooth impaction type identification system described in an embodiment of the present invention.

[0034] Figure 2 This is a schematic flow chart of a method for identifying the type of impacted wisdom teeth based on oral periapical radiographs based on deep learning according to an embodiment of the present invention.

[0035] Figure 3 It is a schematic diagram of a periapical radiograph according to an embodiment of the present invention.

[0036] Figure 4 This is a wisdom tooth labeled binary image according to an embodiment of the present invention.

[0037] Figure 5 Schematic diagram of the wisdom tooth area image cropping process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below by the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0039] It should be understood by those skilled in the art that unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those skilled in the art to which this invention belongs.

[0040] It should also be understood that terms, such as those defined in commonly used dictionaries, should be understood to have a meaning consistent with that in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless as defined herein.

[0041] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements and / or groups thereof.

[0042] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0043] To facilitate understanding of the present invention, the present invention is further explained below with reference to specific embodiments in conjunction with the accompanying drawings, and the specific embodiments do not constitute a limitation on the embodiments of the present invention.

[0044] Those skilled in the art should understand that the drawings are merely schematic diagrams of embodiments, and the components in the drawings are not necessarily necessary for implementing the present invention.

[0045] Example 1

[0046] like Figure 1 As shown, this embodiment 1 provides a system for identifying the type of impacted wisdom teeth based on oral periapical radiographs based on deep learning, and the system includes:

[0047] An acquisition module, used for acquiring an oral periapical radiograph image to be identified;

[0048] A segmentation module is used to process the acquired oral periapical radiograph image to be identified using the trained wisdom tooth segmentation model to obtain a segmented image;

[0049] A cropping module is used to crop the wisdom tooth region of the segmented image to obtain a wisdom tooth region image;

[0050] The recognition module is used to process the wisdom tooth area image using the trained impaction type recognition model to obtain the wisdom tooth impaction type recognition result.

[0051] In this embodiment 1, a method for identifying the type of impacted wisdom teeth is implemented using the above-mentioned system for identifying the type of impacted wisdom teeth, including:

[0052] The acquisition module is used to acquire the oral periapical film image to be identified; the segmentation module is used to process the acquired oral periapical film image to be identified based on the trained wisdom tooth segmentation model to obtain a segmented image; the cropping module is used to crop the wisdom tooth area of ​​the segmented image to obtain a wisdom tooth area image; the recognition module is used to process the wisdom tooth area image based on the trained impaction type recognition model to obtain the wisdom tooth impaction type recognition result.

[0053] In this embodiment 1, training the wisdom tooth segmentation model includes: performing feature enhancement preprocessing on the tooth area in multiple oral periapical film images; marking the wisdom teeth and the types of wisdom tooth impaction in the images after the preprocessing; normalizing the size of the labeled image data, dividing it into a training set, a validation set and a test set, training the basic network and optimizing the parameters to obtain the final trained wisdom tooth segmentation model.

[0054] In this embodiment 1, training the impaction type recognition model includes: based on the segmentation result of the wisdom tooth segmentation model, cropping the wisdom tooth area of ​​the data set image to obtain the wisdom tooth area image; normalizing the size of the cropped wisdom tooth area image data set with the wisdom tooth impaction type annotation, dividing it into a training set, a verification set and a test set, training the basic network and optimizing the parameters to obtain the final trained impaction type recognition model.

[0055] Among them, gray value normalization is used to perform feature enhancement preprocessing on the tooth area in the oral periapical radiograph. The impaction type labels of wisdom teeth include mesial impaction, distal impaction, horizontal impaction, vertical impaction and inverted impaction.

[0056] In this embodiment 1, the wisdom tooth segmentation model is recorded as WTS-Net, and U-Net is used as the network structure of WTS-Net. The wisdom tooth impaction type identification model is recorded as WTC-Net, and ResNet-50 is used as the network structure of WTC-Net.

[0057] Example 2

[0058] like Figure 2 As shown, in this embodiment 2, a method for identifying the type of wisdom tooth impaction based on oral periapical film images based on deep learning is provided, firstly, an oral periapical film image to be identified is obtained; then, the acquired oral periapical film image to be identified is processed using a trained wisdom tooth segmentation model to obtain a segmented image; then, the segmented image is cropped for the wisdom tooth area to obtain a wisdom tooth area image; finally, the wisdom tooth area image is processed using the trained impaction type identification model to obtain a wisdom tooth impaction type identification result.

[0059] Specifically, in this embodiment 2, the oral periapical film image is first preprocessed, the wisdom teeth in the preprocessed image are annotated at the pixel level to obtain category labels, and the training set, validation set and test set are divided; then the wisdom tooth segmentation model (WTS-Net) is trained using the image and category labels, and the model output result is binarized by setting the confidence threshold of the pixel category, and finally the pixel-level classification information of the wisdom tooth is obtained to complete the wisdom tooth segmentation; then, image cropping processing is performed based on the segmentation result of the wisdom tooth to obtain an image containing the wisdom tooth area, and the wisdom tooth area image in the training set is annotated with the wisdom tooth impaction category at the image level; finally, the wisdom tooth impaction type recognition model (WTC-Net) is trained using the wisdom tooth area image and the impaction category annotation to obtain the impaction type classification information of the wisdom tooth.

[0060] In this embodiment 2, the method for identifying the type of impacted wisdom teeth based on oral periapical radiographs based on deep learning specifically includes the following steps when training the wisdom tooth segmentation model and the impacted type recognition model:

[0061] Step 1: Image preprocessing: In order to eliminate the image grayscale distribution deviation caused by different models of equipment, and at the same time enhance the features of the tooth area in the image and improve the image contrast, the preprocessing method mainly uses grayscale normalization, that is, normalizing the image grayscale value to 0-255, as shown in the following formula, where X represents the original grayscale value of the image, and Y represents the grayscale value after normalization:

[0062] Y=(XX min ) / (X max -X min )*255.

[0063] Step 2: Image annotation: After preprocessing, Figure 3 The image shown in the figure is annotated, and the annotated object is wisdom teeth. The annotation content includes: a) binary annotation of each pixel to form a binary image, such as Figure 4 As shown, it is used for training the wisdom tooth segmentation model; b) wisdom tooth impaction type labels, including 5 categories: mesial impaction, distal impaction, horizontal impaction, vertical impaction and inverted impaction, which are used for training the wisdom tooth impaction type recognition model.

[0064] Step 3: Training of wisdom tooth segmentation model: The wisdom tooth segmentation model is recorded as WTS-Net, and U-Net is used as the network structure of WTS-Net. The labeled image data is normalized to 448×448, and then divided into training set, validation set and test set according to 8:1:1. The Adam optimizer is used for model training, the initial learning rate is set to 0.0003, the loss function is Dice Loss, and the number of training iterations is set to 50. After training and parameter optimization, the Dice value of wisdom tooth segmentation on 100 test samples is 0.9431.

[0065] Step 4: Wisdom tooth area image cropping: Based on the segmentation results of the WTS-Net model, the wisdom tooth area is cropped for the dataset image to obtain the wisdom tooth area image. The cropping rules are as follows: Generate a bounding box using the wisdom tooth segmentation binary image, then expand the bounding box to 1 / 3 of the long side of the original bounding box. Finally, use the expanded bounding box to crop the original image and the wisdom tooth segmentation binary image to obtain a dual-channel wisdom tooth area image. The cropping process is as follows: Figure 5 shown.

[0066] Step 5: Training of the model for identifying the impacted wisdom teeth: The model for identifying the impacted wisdom teeth is denoted as WTC-Net, and ResNet-50 is used as the network structure of WTC-Net. The cropped wisdom tooth area image dataset with the impacted wisdom teeth type annotations was normalized to 384×384, and then divided into training set, validation set, and test set according to 8:1:1. The Adam optimizer was used for model training, the initial learning rate was set to 0.0001, the loss function used the cross-entropy loss function, and the number of training iterations was set to 40. After training and parameter optimization, the accuracy of identifying the impacted wisdom teeth on 100 test samples was 0.93.

[0067] Example 3

[0068] Embodiment 3 of the present invention provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the method for identifying the type of impacted wisdom teeth as described above is implemented. The method includes:

[0069] Acquire an oral periapical radiograph image to be identified;

[0070] The trained wisdom tooth segmentation model is used to process the acquired oral periapical radiograph image to be identified to obtain a segmented image;

[0071] Crop the wisdom tooth region of the segmented image to obtain a wisdom tooth region image;

[0072] The trained impaction type recognition model is used to process the wisdom tooth area image to obtain the wisdom tooth impaction type recognition result.

[0073] Example 4

[0074] Embodiment 4 of the present invention provides a computer program (product), including a computer program, wherein when the computer program is executed on one or more processors, the computer program is used to implement the above-mentioned method for identifying the type of impacted wisdom tooth, the method comprising:

[0075] Acquire an oral periapical radiograph image to be identified;

[0076] The trained wisdom tooth segmentation model is used to process the acquired oral periapical radiograph image to be identified to obtain a segmented image;

[0077] Crop the wisdom tooth region of the segmented image to obtain a wisdom tooth region image;

[0078] The trained impaction type recognition model is used to process the wisdom tooth area image to obtain the wisdom tooth impaction type recognition result.

[0079] Example 5

[0080] Embodiment 5 of the present invention provides an electronic device, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes instructions for implementing the above-mentioned method for identifying the type of impacted wisdom teeth, the method comprising:

[0081] Acquire an oral periapical radiograph image to be identified;

[0082] The trained wisdom tooth segmentation model is used to process the acquired oral periapical radiograph image to be identified to obtain a segmented image;

[0083] Crop the wisdom tooth region of the segmented image to obtain a wisdom tooth region image;

[0084] The trained impaction type recognition model is used to process the wisdom tooth area image to obtain the wisdom tooth impaction type recognition result.

[0085] In summary, the method and system for identifying the type of impacted wisdom teeth based on deep learning in oral periapical radiographs provided in the embodiments of the present invention utilize a two-stage deep learning method to accurately identify the type of impacted wisdom teeth; for the impacted wisdom tooth category recognition model WTC-Net, multi-channel input is used to improve classification accuracy.

[0086] Among them, the basic network for training the WTS-Net model can be replaced with any semantic segmentation (such as FCN, DeepLab, U-Net++, etc.) or target detection model (such as YOLO, SSD, Fast R-CNN, Faster R-CNN, etc.); the basic network for training the WTC-Net model can be replaced with any classification model (such as VGG, ResNet, ResNeXt, DenseNet, EfficientNet, RegNet, etc.); the Adam optimizer can be replaced with other optimizers (such as SGD, AdamW, etc.).

[0087] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0088] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0089] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0090] These computer program instructions can also be loaded onto a computer or other programmable data processing device, and a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0091] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative work on the basis of the technical solution disclosed in the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for identifying the type of impacted wisdom teeth, characterized in that: include: Acquire an oral periapical radiograph image to be identified; The trained wisdom tooth segmentation model is used to process the acquired oral periapical radiograph image to be identified to obtain a segmented image; Training the wisdom tooth segmentation model includes: performing feature enhancement preprocessing on the tooth area in multiple oral periapical film images; marking the wisdom teeth and the types of wisdom tooth impaction on the preprocessed images; normalizing the size of the labeled image data, dividing them into training sets, verification sets and test sets, training the basic network and optimizing the parameters to obtain the final trained wisdom tooth segmentation model; when marking the wisdom teeth, specifically: binary labeling pixel by pixel to form a binary image; the wisdom tooth impaction type label includes mesial impaction, distal impaction, horizontal impaction, vertical impaction and inverted impaction; Crop the wisdom tooth region of the segmented image to obtain a wisdom tooth region image; The trained impaction type recognition model is used to process the wisdom tooth area image to obtain the wisdom tooth impaction type recognition result; training the impaction type recognition model includes: based on the segmentation result of the wisdom tooth segmentation model, the wisdom tooth area is cropped on the data set image to obtain the wisdom tooth area image; the cropped wisdom tooth area image data set with the wisdom tooth impaction type annotation is normalized, divided into a training set, a validation set and a test set, the basic network is trained and the parameters are optimized to obtain the final trained impaction type recognition model; when the wisdom tooth area is cropped on the data set image, specifically: a bounding box is generated using the wisdom tooth segmentation binary image, and then the bounding box is expanded, and the expansion length is 1 / 3 of the long side of the original bounding box. Finally, the expanded bounding box is used to crop the original image and the wisdom tooth segmentation binary image to obtain a dual-channel wisdom tooth area image.

2. The method for identifying the type of impacted wisdom tooth according to claim 1, characterized in that: Gray value normalization is used to perform feature enhancement preprocessing on the tooth area in oral periapical radiographs.

3. The method for identifying the type of impacted wisdom tooth according to claim 1, characterized in that: The basic network used to train the wisdom tooth segmentation model is: One of the semantic segmentation networks FCN, DeepLab or U-Net++; or, One of the object detection networks YOLO, SSD, Fast R-CNN or Faster R-CNN.

4. The method for identifying the type of impacted wisdom tooth according to claim 1, characterized in that: The basic network for training the impaction type recognition model is one of VGG, ResNet, ResNeXt, DenseNet, EfficientNet or RegNet.

5. A system for identifying the type of impacted wisdom teeth, characterized in that: include: An acquisition module, used for acquiring an oral periapical radiograph image to be identified; A segmentation module is used to process the acquired oral periapical radiograph image to be identified using the trained wisdom tooth segmentation model to obtain a segmented image; Training the wisdom tooth segmentation model includes: performing feature enhancement preprocessing on the tooth area in multiple oral periapical film images; marking the wisdom teeth and the types of wisdom tooth impaction on the preprocessed images; normalizing the size of the labeled image data, dividing them into training sets, verification sets and test sets, training the basic network and optimizing the parameters to obtain the final trained wisdom tooth segmentation model; when marking the wisdom teeth, specifically: binary labeling pixel by pixel to form a binary image; the wisdom tooth impaction type label includes mesial impaction, distal impaction, horizontal impaction, vertical impaction and inverted impaction; A cropping module is used to crop the wisdom tooth region of the segmented image to obtain a wisdom tooth region image; The recognition module is used to process the wisdom tooth area image using the trained impaction type recognition model to obtain the wisdom tooth impaction type recognition result; training the impaction type recognition model includes: based on the segmentation result of the wisdom tooth segmentation model, the wisdom tooth area is cropped on the data set image to obtain the wisdom tooth area image; the size of the cropped wisdom tooth area image data set with the wisdom tooth impaction type annotation is normalized, divided into a training set, a verification set and a test set, and the basic network is trained and the parameters are optimized to obtain the final trained impaction type recognition model; when the wisdom tooth area is cropped on the data set image, specifically: a bounding box is generated using the wisdom tooth segmentation binary image, and then the bounding box is expanded, and the expansion length is 1 / 3 of the long side of the original bounding box. Finally, the expanded bounding box is used to crop the original image and the wisdom tooth segmentation binary image to obtain a dual-channel wisdom tooth area image.

6. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by the processor, the method for identifying the type of impacted wisdom teeth as described in any one of claims 1-4 is implemented.

7. An electronic device, characterized in that: include: A processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes instructions for implementing the method for identifying the type of impacted wisdom teeth as described in any one of claims 1-4.

Images