Processing method and processing device of intraoral scanning image, terminal and computer storage medium
Through improved lightweight segmentation network architecture and data expansion technology, the image segmentation model is optimized, and the applicability problem of intraoral scanning image processing method is solved, efficient and accurate soft tissue removal is achieved, and it is suitable for intraoral scanning image processing in various scenarios.
Patent Information
- Application Number
- CN202411950402.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-27
- Publication Date
- 2025-06-06
AI Technical Summary
In the prior art, the processing method of intraoral scanning images is difficult to apply to all scenarios, resulting in difficult and costly processing, and is not suitable for real-time needs of clinical diagnosis.
The lightweight segmentation network architecture (such as the improved YOLO V8 segmentation network) is adopted to combine data expansion and distillation technology to optimize the image segmentation model by training and testing the data sets to achieve efficient segmentation of scanned images in the mouth.
It realizes efficient and accurate removal of soft tissue areas in the intraoral scanning image in a short period of time, which is suitable for all kinds of scenarios and assists doctors in rapid diagnosis.
Smart Images
Figure CN120107280A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of image processing and relates to an image segmentation technology, and in particular to a method, a processing device, a terminal and a computer storage medium for processing intraoral scanned images. Background Art
[0002] Intraoral scan images are RGB images obtained by scanning the patient's oral cavity using digital technology, which are used to assist doctors in obtaining the patient's dental information. In actual application scenarios, in order to intuitively and clearly reflect the patient's dental information, it is usually necessary to remove the soft tissue part of the intraoral scan image.
[0003] At present, the method for removing soft tissue parts on intraoral scan images is usually to use image segmentation methods to identify and segment the soft tissue parts on the intraoral scan images for removal. Specifically, the image segmentation method can be an image processing method such as threshold segmentation, edge detection or region growing. However, since oral scenes are usually more complex, a single image segmentation method is difficult to apply to intraoral scan images in all scenes, and using multiple image segmentation methods to process intraoral scan images will increase the difficulty and cost of image processing, which is not suitable for actual application scenarios.
[0004] Therefore, how to obtain an image processing method to cover intraoral scan images in all scenarios is an urgent problem to be solved by those skilled in the art. Summary of the invention
[0005] The purpose of the present application is to provide a method, a processing device, a terminal and a computer storage medium for processing intraoral scanned images, so as to solve the problem that the processing method of intraoral scanned images in the prior art cannot be applied to intraoral scanned images in all scenarios, or the processing is difficult and costly.
[0006] In a first aspect, the present application provides a method for processing an intraoral scan image, comprising:
[0007] Acquire a plurality of intraoral scan image data, pre-process each of the intraoral scan image data to mark the area to be segmented, and randomly divide the data into a training data set and a test data set;
[0008] Based on the training data set, a lightweight segmentation network architecture is trained to obtain an image segmentation model to be tested, and the image segmentation model to be tested is tested using the test data set. For the image segmentation model that passes the test, an intraoral scan image segmentation model is obtained based on the image segmentation model that passes the test;
[0009] Based on the intraoral scan image segmentation model, image segmentation is performed on the data to be processed, and the corresponding intraoral scan image processing results are output.
[0010] In one embodiment of the present application, the lightweight segmentation network architecture is improved by the YOLO V8 segmentation network architecture, including: replacing the Backbone module in the YOLO V8 segmentation network architecture with ShuffleNet to obtain the lightweight segmentation network architecture.
[0011] In one embodiment of the present application, the image segmentation model to be tested is tested using the test data set, and for the image segmentation model that fails the test, the training data set is augmented to obtain an augmented training data set; based on the augmented training data set, the lightweight segmentation network architecture is retrained to obtain a new image segmentation model to be tested, and based on the new image segmentation model to be tested, the test process is re-executed using the test data set.
[0012] In one embodiment of the present application, the step of performing data expansion on the training data set to obtain an expanded training data set includes:
[0013] Performing nonlinear enhancement processing on each of the intraoral scanned image data in the training data set to obtain corresponding enhanced image data;
[0014] and / or, performing data synthesis processing on each of the intraoral scan image data in the training data set to obtain corresponding synthesized image data;
[0015] The expanded training data set is generated based on each of the intraoral scan image data in combination with each of the enhanced image data and / or each of the synthesized image data.
[0016] In one embodiment of the present application, the performing nonlinear enhancement processing on each of the intra-oral scanned image data in the training data set to obtain the corresponding enhanced image data includes: performing at least one image processing process of flipping, mirroring, rotating, translating, scaling and gamma transform on each of the intra-oral scanned image data in the training data set to obtain the corresponding enhanced image data;
[0017] The step of performing data synthesis processing on each of the intraoral scan image data in the training data set to obtain corresponding synthesized image data includes: based on each of the intraoral scan image data in the training data set, outputting corresponding synthesized image data by generating an adversarial network.
[0018] In one embodiment of the present application, the method of obtaining an intraoral scan image segmentation model based on a tested image segmentation model comprises: performing a distillation process on the tested image segmentation model to obtain a distillation model; and obtaining the intraoral scan image segmentation model based on the distillation model.
[0019] In one embodiment of the present application, the step of obtaining the intra-oral scan image segmentation model based on the distillation model includes: performing a quantization process on the distillation model to obtain the intra-oral scan image segmentation model.
[0020] In a second aspect, the present application provides a device for processing intraoral scanned images, including a preprocessing module, a model building module, and an image segmentation module;
[0021] The preprocessing module is used to obtain a plurality of intraoral scan image data, preprocess each of the intraoral scan image data to mark the area to be segmented, and randomly divide the data into a training data set and a test data set;
[0022] The model building module is used to train a lightweight segmentation network architecture based on the training data set to obtain an image segmentation model to be tested, and use the test data set to test the image segmentation model to be tested. For the image segmentation model that passes the test, an intraoral scan image segmentation model is obtained based on the image segmentation model that passes the test;
[0023] The image segmentation module is used to perform image segmentation on the data to be processed based on the intraoral scan image segmentation model, and output the corresponding intraoral scan image processing results.
[0024] In a third aspect, the present application provides a terminal, including: a processor and a memory, wherein the memory is communicatively connected to the processor;
[0025] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes the method for processing intraoral scanned images as described above.
[0026] In a fourth aspect, the present application provides a computer storage medium storing a computer program, which, when executed by a processor, implements the method for processing intraoral scanned images as described above.
[0027] As described above, the present application provides a method, a processing device, a terminal and a computer storage medium for processing intraoral scanned images. Through an intraoral scanned image segmentation model, image segmentation is performed on the data to be processed to eliminate the soft tissue area on the data to be processed, so as to assist doctors in diagnosis. The image segmentation has high accuracy and short processing time, and is suitable for various types of images to be processed in various scenarios, and has high industrial application value. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 Shown is a flowchart of a method for processing an intraoral scan image described in an embodiment of the present application.
[0029] Figure 2 Shown is a schematic diagram of the recognition of the intraoral scan image by the intraoral scan image segmentation model described in an embodiment of the present application.
[0030] Figure 3 Shown is a flow chart of a method for acquiring a segmentation model of an image to be tested described in an embodiment of the present application.
[0031] Figure 4 Shown is a flow chart of a method for obtaining an expanded training data set described in an embodiment of the present application.
[0032] Figure 5 Shown is a schematic diagram of the structure of a device for processing intraoral scanned images described in an embodiment of the present application.
[0033] Figure 6 Shown is a schematic diagram of the structure of a terminal described in an embodiment of the present application.
[0034] Description of Reference Numerals
[0035] 31 Preprocessing Module
[0036] 32 Model building modules
[0037] 33 Image segmentation module
[0038] 40Terminal
[0039] 41 processors
[0040] 42 Memory
[0041] 43 User Interface
[0042] 44 network interfaces
[0043] 45 bus system
[0044] 421 Operating System
[0045] 422 Application DETAILED DESCRIPTION
[0046] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.
[0047] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present application, and thus the drawings only show components related to the present application rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed at will, and the component layout may also be more complicated.
[0048] The existing methods for processing intraoral scanned images often segment the intraoral scanned images through image processing methods such as threshold segmentation, edge detection or region growing. However, due to the complexity of the scene in the oral cavity, a single image processing method cannot be applied to all images to be processed. For example, images with uneven illumination or uneven grayscale distribution do not have good processing effects using the threshold segmentation method, and images with unclear edges or large noise interference cannot be processed using the edge detection method. Although better image processing effects can be obtained by obtaining corresponding image processing methods based on the characteristics of each image to be processed for segmentation, this also leads to greater difficulty in image processing, higher time costs, and lower image processing efficiency. When the intraoral scanned image processing method is applied to clinical treatment, it is usually necessary to complete image processing in a relatively short time to assist doctors in diagnosis in real time. Therefore, this processing method is not suitable for actual application scenarios.
[0049] In response to the technical problems existing in the prior art, the following embodiments of the present application provide a method, a processing device, a terminal and a computer storage medium for processing intraoral scan images. Through an intraoral scan image segmentation model, the image to be processed is segmented to obtain the processing result of the intraoral scan image, thereby achieving accurate and efficient image processing, and then better assisting doctors in diagnosis, which is conducive to the actual clinical application of the intraoral scan image processing method.
[0050] The following embodiments of the present application provide a method, a processing device, a terminal and a computer storage medium for processing intraoral scanned images, including but not limited to image segmentation applied to intraoral scanned images. The following will describe the soft tissue removal processing of intraoral scanned images as an example.
[0051] The technical solutions in the embodiments of the present application will be described in detail below in conjunction with the drawings in the embodiments of the present application.
[0052] like Figure 1 As shown, this embodiment provides a method for processing an intraoral scan image, comprising:
[0053] S100, acquiring a plurality of intraoral scan image data, preprocessing each of the intraoral scan image data to mark a to-be-segmented area, and randomly dividing the data into a training data set and a test data set.
[0054] Wherein, the preprocessing includes image marking to mark the area to be segmented on the intraoral scan image data.
[0055] Specifically, the soft tissue area to be removed on the intraoral scanned image data can be manually outlined, or the soft tissue area to be removed on the intraoral scanned image data can be outlined by marking software. Exemplarily, the grayscale value of part of the outlined soft tissue area is set to 0, and the grayscale value of other areas is set to 1, so as to facilitate subsequent model training and testing.
[0056] The training data set includes the intraoral scan image data used for model training, and the test data set includes the intraoral scan image data used for model testing. It should be noted that the training data set and the test data set are randomly divided, that is, the training data set and the test data set both contain data of various scenes and types, so as to improve the accuracy of model training and test results.
[0057] S200, based on the training data set, training a lightweight segmentation network architecture to obtain an image segmentation model to be tested, and using the test data set to test the image segmentation model to be tested, and for the image segmentation model that passes the test, obtaining an intraoral scan image segmentation model based on the image segmentation model that passes the test.
[0058] Among them, the lightweight segmentation network architecture is a framework of a network model for detecting and segmenting target areas in image data. By performing deep learning on the data in the training data set, a network model for detecting and segmenting soft tissue areas in the intraoral scan image data is obtained.
[0059] Further, in order to improve the image processing efficiency of the model, the lightweight segmentation network architecture described in this embodiment is a network architecture with a lightweight design. Exemplarily, the lightweight segmentation network architecture is improved based on the lightweight design by the YOLO V8 segmentation network architecture. Specifically, the Backbone module in the YOLO V8 segmentation network architecture is replaced with ShuffleNet to obtain the lightweight segmentation network architecture. It should be noted that those skilled in the art should be aware of the specific composition and architecture of the YOLO V8 segmentation network architecture, wherein the Backbone module is used to extract the target to be detected from the image data, and the Backbone module is replaced with ShuffleNet, thereby convolving the channels to reduce the amount of calculation, thereby improving the image processing efficiency of the acquired intraoral scan image segmentation model.
[0060] The lightweight segmentation network architecture is trained to obtain the image segmentation model to be tested. In order to determine the availability of the image segmentation model, the present embodiment also tests the image segmentation model to be tested through the test training set. Only when the test passes, the accuracy of image processing by the image segmentation model meets clinical requirements, that is, the image segmentation model that passes the test is used as the intraoral scanning image segmentation model.
[0061] The image segmentation model to be tested passes the test only when it can segment all soft tissue regions for any of the intraoral scanned image data in the test data set. Based on this, the intraoral scanned image segmentation model obtained in this embodiment has a high accuracy, and when used in actual clinical applications, it can achieve a good intraoral scanned image processing effect, and has a high industrial application value.
[0062] like Figure 2 As shown in FIG. 1 , an example of the recognition of the intraoral scan image by the intraoral scan image segmentation model is shown, wherein the left image is the original image a of the intraoral scan image, and the right image is the recognition result b of the intraoral scan image. The black area is the area used for segmentation on the intraoral scan image, that is, the recognized soft tissue area. Figure 2 It can be seen that the intraoral scan image segmentation model obtained in this embodiment can identify all soft tissue areas on the intraoral scan image. Based on the recognition result, it is possible to eliminate part of all soft tissue areas on the intraoral scan image, that is, to achieve highly accurate intraoral scan image segmentation processing to assist doctors in diagnosis.
[0063] In some optional embodiments, in order to further improve the image processing efficiency of the intraoral scan image segmentation model, after the image segmentation model test passes, a distillation process is performed on it to obtain a distillation model; based on the distillation model, the intraoral scan image segmentation model is obtained.
[0064] Specifically, the tested image segmentation model is used as a teacher model, and a corresponding student model is designed, wherein the network structure of the student model is simpler and has fewer parameters; based on the processing result of the teacher model on the intraoral scan image data, the difference between the output of the teacher model and the output of the student model is calculated to construct a loss function; based on the loss function, the student model is updated and optimized so that the processing result of the student model on the intraoral scan image data approaches the teacher model, and the student model is output as the distillation model to realize the distillation of the model to be tested, thereby further improving the image processing efficiency of the intraoral scan image segmentation model, thereby obtaining image processing results faster and adapting to actual clinical application needs.
[0065] Furthermore, this embodiment performs a quantization process on the distillation model, and uses the quantized model as the intraoral scanning image segmentation model to further simplify the structure and parameters of the intraoral scanning image segmentation model, thereby effectively improving the image processing efficiency of the intraoral scanning image segmentation model.
[0066] Specifically, the parameters in the distillation model are converted into a representation with lower precision, for example, the representation of each parameter in the distillation model is converted from a floating point number to an integer.
[0067] Based on this, this embodiment trains the lightweight segmentation network architecture through the training data set to obtain the image segmentation model to be tested, and determines the availability of the image segmentation model through the test data set to ensure the accuracy of the acquired intraoral scan image segmentation model. Furthermore, this embodiment further improves the image processing efficiency of the intraoral scan image segmentation model by performing a distillation process and a quantization process on the image segmentation model that passes the test. Specifically, the image processing process can be completed within 5ms, with a fast processing speed, so as to be more adapted to actual clinical application requirements, and is conducive to the actual industrial application of the intraoral scan image processing method.
[0068] S300, performing image segmentation on the data to be processed based on the intraoral scan image segmentation model, and outputting the corresponding intraoral scan image processing result.
[0069] The data to be processed is the intraoral RGB image to be processed. For example, in a clinical application scenario, it is the intraoral RGB image obtained by scanning the patient's oral cavity through digital technology.
[0070] Specifically, the image data to be processed is acquired through digital technology or the like, and the image data to be processed is segmented through the intraoral scan image segmentation model to output the corresponding intraoral scan image processing result, that is, the intraoral RGB image after removing the soft tissue area, so as to assist the doctor in diagnosis.
[0071] It should be noted that in step S200, the image segmentation model to be tested is tested using the test data set. If the test fails, the accuracy of the image segmentation model cannot meet the actual application requirements, and the image segmentation model to be tested is re-acquired. Specifically, in some optional implementations, such as Figure 3 As shown, re-obtaining the image segmentation model to be tested includes:
[0072] S210: Perform data expansion on the training data set to obtain an expanded training data set.
[0073] By expanding the training data set, the amount of data used to train the lightweight segmentation network architecture is increased, thereby improving the accuracy of the acquired image segmentation model to be tested to meet clinical needs.
[0074] In some optional embodiments, such as Figure 4 As shown, the data expansion of the training data set to obtain an expanded training data set includes:
[0075] S211, performing nonlinear enhancement processing on each of the intra-oral scan image data in the training data set, and / or performing data synthesis processing on each of the intra-oral scan image data in the training data set.
[0076] Wherein, each of the intraoral scanned image data in the training data set is subjected to nonlinear enhancement processing to obtain corresponding enhanced image data. The nonlinear enhancement processing refers to performing nonlinear transformation on the intraoral scanned image data to enhance the diversity of data in the training data set.
[0077] Specifically, by changing the brightness, image size or resolution of each of the intraoral scan image data to simulate different acquisition conditions of the intraoral scan image data, the training data set includes data of various types in various scenes, so as to improve the training accuracy of the lightweight segmentation network architecture, and further improve the accuracy of the intraoral scan image segmentation model.
[0078] Exemplarily, the nonlinear enhancement processing includes but is not limited to image processing processes of flipping, mirroring, rotation, translation, scaling or gamma transform, that is, at least one image processing process of flipping, mirroring, rotation, translation, scaling and gamma transform is performed on each of the intraoral scan image data in the training data set to obtain corresponding enhanced image data.
[0079] Data synthesis processing is performed on each of the intraoral scan image data in the training data set to obtain corresponding synthesized image data.
[0080] Among them, the data synthesis processing refers to customizing and generating new intraoral scan image data based on each of the intraoral scan image data through a generative adversarial network or other means, that is, based on each of the intraoral scan image data in the training data set, the corresponding synthetic image data is output through a generative adversarial network.
[0081] Based on this, through the linear enhancement processing and the data synthesis processing, this embodiment obtains and expands the data of the intraoral scan image data, and by simulating different acquisition conditions of the intraoral scan image data, the training data set covers various types of data in various scenes, thereby improving the accuracy of the intraoral scan image segmentation model obtained by training.
[0082] S212: Generate the expanded training data set based on each of the intraoral scan image data in combination with each of the enhanced image data and / or each of the synthesized image data.
[0083] Specifically, each of the intraoral scan image data, as well as each of the enhanced image data and / or each of the synthesized image data obtained based on each of the intraoral scan image data, are obtained, and the total set of these data is used as an expanded training data set for subsequent model training.
[0084] S220: Based on the expanded training data set, retrain the lightweight segmentation network architecture to obtain a new image segmentation model to be tested.
[0085] Specifically, deep learning is performed on the data in the expanded training data set to train the lightweight segmentation network architecture, thereby obtaining a trained network model for detecting and segmenting soft tissue areas in the intraoral scan image data.
[0086] Since the expanded training data set contains more image data after expansion, the image segmentation model to be tested obtained by training based on the expanded training data set has higher accuracy, so that the image segmentation model to be tested can pass the test.
[0087] Furthermore, in step S200, the lightweight segmentation network architecture is trained based on the training data set to obtain the image segmentation model to be tested, and the training data set may also be expanded, so that during the process of training the lightweight segmentation network architecture, the training data set has a large amount of data covering various scenes and types, and the image segmentation model to be tested obtained by training based on the expanded training data set has higher accuracy, thereby improving the accuracy of the intraoral scan image segmentation model, and then improving the processing effect of the intraoral scan image, which is conducive to better assisting doctors in diagnosis.
[0088] S230, re-executing the test process using the test data set based on the new image segmentation model to be tested.
[0089] The new image segmentation model to be tested is retested to determine whether it meets the actual application requirements, so that the intraoral scan image segmentation model obtained based on the image segmentation model to be tested has a higher accuracy, so as to facilitate the application of the intraoral scan image processing method provided in this embodiment in actual clinical scenarios.
[0090] Specifically, for the testing process and principle of the new image segmentation model to be tested, please refer to the above content, and this application will not go into details here.
[0091] like Figure 5 As shown, the present embodiment provides a device for processing intraoral scanned images, comprising a preprocessing module 31 , a model building module 32 and an image segmentation module 33 .
[0092] The preprocessing module 31 is used to obtain a plurality of intraoral scan image data, preprocess each of the intraoral scan image data to mark the area to be segmented, and randomly divide the data into a training data set and a test data set;
[0093] The model building module 32 is used to train a lightweight segmentation network architecture based on the training data set to obtain an image segmentation model to be tested, and use the test data set to test the image segmentation model to be tested. For the image segmentation model that passes the test, an intraoral scan image segmentation model is obtained based on the image segmentation model that passes the test;
[0094] The image segmentation module 33 is used to perform image segmentation on the data to be processed based on the intraoral scan image segmentation model, and output the corresponding intraoral scan image processing results.
[0095] Based on the same technical concept, the method for processing the intraoral scanned image provided in the embodiment of the present invention can be implemented on the terminal side or the server side.
[0096] like Figure 6 As shown, it is an optional hardware structure diagram of a terminal provided by an embodiment of the present invention, and the terminal can be a mobile phone, a computer device, a tablet device, a personal digital processing device, a factory background processing device, etc. The terminal includes: at least one processor 41, a memory 42, at least one network interface 44 and a user interface 43. The various components in the device are coupled together through a bus system 45. It can be understood that the bus system 45 is used to realize the connection and communication between these components. In addition to the data bus, the bus system also includes a power bus, a control bus and a status signal bus.
[0097] The user interface 43 may include a display, a keyboard, a mouse, a trackball, a click gun, keys, buttons, a touch pad or a touch screen.
[0098] It is understood that the memory 42 can be a volatile memory or a non-volatile memory, and can also include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM). The memory characterized by the embodiments of the present invention is intended to include but is not limited to these and any other suitable categories of memory.
[0099] The memory 42 in the embodiment of the present invention is used to store various categories of data to support the operation of the terminal. Examples of these data include: any executable program used to operate on the terminal 40, such as an operating system 421 and an application 422; the operating system 421 includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application 422 may include various applications, such as a media player (MediaPlayer), a browser (Browser), etc., for implementing various application services. The method for processing the intraoral scanned image provided in the embodiment of the present invention may be included in the application 422.
[0100] The method disclosed in the above embodiment of the present invention can be applied to the processor 41, or implemented by the processor 41. The processor 41 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit in the processor 41 or the instruction in the form of software. The above processor may be a general processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The processor 41 can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiment of the present invention. The general processor 41 may be a microprocessor or any conventional processor, etc. In combination with the steps of the accessory optimization method provided in the embodiment of the present invention, it can be directly embodied as a hardware decoding processor to execute, or it can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium, which is located in a memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.
[0101] In an exemplary embodiment, the terminal 40 may be one or more application specific integrated circuits (ASIC), DSP, programmable logic device (PLD), complex programmable logic device (CPLD) to execute the aforementioned method.
[0102] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is called by a processor, the method for processing the intraoral scanned image provided by the present invention is implemented.
[0103] Among them, the computer-readable storage medium can be a tangible device that can hold and store instructions used by the instruction execution device. The computer-readable storage medium can be, for example, (but not limited to) an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the above. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, and a mechanical encoding device.
[0104] The computer-readable program characterized herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. A network adapter card or network interface in each computing / processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing / processing device.
[0105] In summary, the present application uses the intraoral scan image segmentation model to efficiently and accurately realize image segmentation of the image to be processed, and is suitable for various types of images to be processed in various scenarios, which is conducive to the actual clinical application of the intraoral scan image processing method and has high industrial application value.
[0106] The descriptions of the processes or structures corresponding to the above-mentioned figures have different emphases. For parts that are not described in detail in a certain process or structure, please refer to the relevant descriptions of other processes or structures.
[0107] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.
Claims
1. A method for processing an intraoral scan image, comprising: Acquire a plurality of intraoral scan image data, pre-process each of the intraoral scan image data to mark the area to be segmented, and randomly divide the data into a training data set and a test data set; Based on the training data set, a lightweight segmentation network architecture is trained to obtain an image segmentation model to be tested, and the image segmentation model to be tested is tested using the test data set. For the image segmentation model that passes the test, an intraoral scan image segmentation model is obtained based on the image segmentation model that passes the test; Based on the intraoral scan image segmentation model, image segmentation is performed on the data to be processed, and the corresponding intraoral scan image processing results are output.
2. The processing method according to claim 1, characterized in that: The lightweight segmentation network architecture is improved by the YOLO V8 segmentation network architecture, including: replacing the Backbone module in the YOLO V8 segmentation network architecture with ShuffleNet to obtain the lightweight segmentation network architecture.
3. The processing method according to claim 1, characterized in that: The image segmentation model to be tested is tested using the test data set. For the image segmentation model that fails the test, the training data set is augmented to obtain an augmented training data set. Based on the augmented training data set, the lightweight segmentation network architecture is retrained to obtain a new image segmentation model to be tested. Based on the new image segmentation model to be tested, the test process is re-executed using the test data set.
4. The processing method according to claim 3, characterized in that: The step of performing data expansion on the training data set to obtain an expanded training data set includes: Performing nonlinear enhancement processing on each of the intraoral scan image data in the training data set to obtain corresponding enhanced image data; and / or, performing data synthesis processing on each of the intraoral scan image data in the training data set to obtain corresponding synthesized image data; The expanded training data set is generated based on each of the intraoral scan image data in combination with each of the enhanced image data and / or each of the synthesized image data.
5. The processing method according to claim 4, characterized in that: The step of performing nonlinear enhancement processing on each of the intraoral scanned image data in the training data set to obtain corresponding enhanced image data comprises: performing at least one image processing process of flipping, mirroring, rotating, translating, scaling and gamma transform on each of the intraoral scanned image data in the training data set to obtain corresponding enhanced image data; The step of performing data synthesis processing on each of the intraoral scan image data in the training data set to obtain corresponding synthesized image data includes: based on each of the intraoral scan image data in the training data set, outputting corresponding synthesized image data by generating an adversarial network.
6. The processing method according to claim 1, characterized in that: The method of obtaining an intraoral scan image segmentation model based on a tested image segmentation model comprises: performing a distillation process on the tested image segmentation model to obtain a distillation model; and obtaining the intraoral scan image segmentation model based on the distillation model.
7. The processing method according to claim 6, characterized in that: The step of obtaining the intra-oral scan image segmentation model based on the distillation model includes: performing a quantization process on the distillation model to obtain the intra-oral scan image segmentation model.
8. A device for processing intraoral scanned images, characterized in that: Includes preprocessing module, model building module and image segmentation module; The preprocessing module is used to obtain a plurality of intraoral scan image data, preprocess each of the intraoral scan image data to mark the area to be segmented, and randomly divide the data into a training data set and a test data set; The model building module is used to train a lightweight segmentation network architecture based on the training data set to obtain an image segmentation model to be tested, and use the test data set to test the image segmentation model to be tested. For the image segmentation model that passes the test, an intraoral scan image segmentation model is obtained based on the image segmentation model that passes the test; The image segmentation module is used to perform image segmentation on the data to be processed based on the intraoral scan image segmentation model, and output the corresponding intraoral scan image processing results.
9. A terminal, characterized in that: include: A processor and a memory, wherein the memory is communicatively connected to the processor; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes the method for processing an intraoral scan image according to any one of claims 1 to 7.
10. A computer storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for processing an intraoral scan image according to any one of claims 1 to 7 is implemented.