Dental image processing method, device, computer equipment and storage medium

Abstract

The present application is applicable to the field of image technology, and in particular, to a dental image processing method, device, computer equipment and storage medium. The method obtains a panoramic dental film to be analyzed, extracts features of the teeth in the panoramic dental film, determines the root length of each tooth and the corresponding tooth, segments the periodontal bone area in the panoramic dental film, obtains the periodontal bone boundary that characterizes each tooth, segments the enamel area in the panoramic dental film, obtains the enamel boundary that characterizes each tooth, determines the bone loss distance between the periodontal bone and enamel of the target tooth based on the periodontal bone boundary and enamel boundary of the target tooth, determines the periodontitis characterization amount of the target tooth based on the bone loss distance and the root length, and performs a stage analysis on the panoramic dental film. By analyzing the tooth length and the periodontal bone and enamel boundaries, the staging of bone loss and periodontitis is automatically calculated, thereby improving the automation and accuracy of the panoramic dental film analysis.

Description

Technical Field

[0001] The present application is applicable to the field of image processing technology, and in particular relates to a dental image processing method, apparatus, computer equipment, and storage medium. Background Art

[0002] Periodontitis is a common dental disease, and its diagnosis and treatment rely on accurate staging of dental lesions. Currently, dental imaging primarily includes periapical images, panoramic X-rays (also known as panoramic dental radiographs), and cone-beam computed tomography (CBCT) images. Periapical images have a narrow field of view and can only show a portion of the tooth, making them suitable for observing single teeth or localized details. Panoramic X-rays display the teeth, bone structure, and surrounding tissues of the entire mouth. Due to their wide field of view and stable image quality, they are ideal for diagnosing widespread lesions such as periodontitis. CBCT images, a three-dimensional imaging technology, provide greater detail but are susceptible to artifacts (interference in the image that does not represent actual structures) and noise, which can reduce diagnostic accuracy.

[0003] Currently, neural network models can be used for automated image classification and processing. However, edge detection and segmentation of panoramic X-ray images can only provide a relatively rough picture of the tooth lesion area. For the assessment of periodontitis staging, accurate staging results cannot be obtained. In addition, due to insufficient data samples, the accuracy of periodontitis staging results obtained by directly training neural networks on images is low. It can be seen that conventional neural network models lack an effective response to data imbalance problems, making it difficult to simultaneously ensure automation and high precision.

[0004] Therefore, how to classify and process panoramic dental radiographs to improve the accuracy of automatic periodontitis staging has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, the embodiments of the present application provide a dental image processing method, apparatus, computer equipment and storage medium to solve the problem of how to classify and process panoramic dental radiographs to improve the accuracy of automatic periodontitis staging.

[0006] In a first aspect, an embodiment of the present application provides a dental image processing method, the dental image processing method comprising: Obtaining a panoramic dental film to be analyzed, performing feature extraction on the teeth in the panoramic dental film, and determining the root length of each tooth and the corresponding tooth; Segmenting the periodontal bone region in the panoramic dental film to obtain a periodontal bone boundary representing each tooth, and segmenting the enamel region in the panoramic dental film to obtain an enamel boundary representing each tooth; Acquire a target tooth, and determine a bone loss distance between the periodontal bone and the enamel of the target tooth according to the periodontal bone boundary and the enamel boundary of the target tooth; The periodontitis characterization value of the target tooth is determined based on the bone loss interval and the tooth root length, and the staging analysis result of the panoramic dental film is determined based on the periodontitis characterization value.

[0007] In a second aspect, an embodiment of the present application provides a dental image processing device, comprising: A tooth feature analysis module is used to obtain a panoramic dental film to be analyzed, extract features of the teeth in the panoramic dental film, and determine the root length of each tooth and the corresponding tooth; a periodontal feature analysis module, configured to segment the periodontal bone region in the panoramic dental film to obtain a periodontal bone boundary representing each tooth, and segment the enamel region in the panoramic dental film to obtain an enamel boundary representing each tooth; a bone loss calculation module, configured to obtain a target tooth and determine a bone loss distance between the periodontal bone and the enamel of the target tooth according to the periodontal bone boundary and the enamel boundary of the target tooth; The periodontal staging analysis module is used to determine the periodontitis characterization value of the target tooth based on the bone loss interval and the tooth root length, and determine the staging analysis result of the panoramic dental film based on the periodontitis characterization value.

[0008] In a third aspect, an embodiment of the present application provides a computer device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the dental image processing method as described in the first aspect is implemented.

[0009] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the dental image processing method as described in the first aspect is implemented.

[0010] Compared with the prior art, the beneficial effects of the embodiments of the present application are as follows: the present application obtains a panoramic dental film to be analyzed, extracts features of the teeth in the panoramic dental film, determines the root length of each tooth and the corresponding tooth, segments the periodontal bone area in the panoramic dental film, obtains the periodontal bone boundary representing each tooth, segments the enamel area in the panoramic dental film, obtains the enamel boundary representing each tooth, obtains the target tooth, determines the bone loss distance between the periodontal bone and enamel of the target tooth based on the periodontal bone boundary and enamel boundary of the target tooth, determines the periodontitis characterization amount of the target tooth based on the bone loss distance and the root length, and determines the staging analysis result of the panoramic dental film based on the periodontitis characterization amount. By analyzing the tooth length and the periodontal bone and enamel boundaries, it is possible to classify the image features of the teeth in the panoramic dental film, automatically calculate the bone loss, and determine the stage of periodontitis in combination with the tooth length, thereby improving the automation and accuracy of the analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0012] Figure 1 This is a schematic diagram of an application environment of a dental image processing method provided in Example 1 of the present application; Figure 2 This is a flow chart of a dental image processing method provided in Example 2 of the present application; Figure 3 This is a schematic diagram of tooth feature extraction provided in Example 2 of the present application; Figure 4 This is a schematic diagram of extracting the periodontal bone boundary and the enamel boundary provided in Example 2 of the present application; Figure 5 This is a schematic diagram of marking bone loss spacing provided in Example 2 of the present application; Figure 6 This is a schematic diagram of the principle of a dental image processing method provided in Example 2 of the present application; Figure 7 This is a flow chart of a dental image processing method provided in Example 3 of the present application; Figure 8 This is a flow chart of a dental image processing method provided in Example 4 of the present application; Figure 9 This is a flow chart of a dental image processing method provided in Example 5 of the present application; Figure 10 This is a schematic structural diagram of a dental image processing device provided in Example 6 of the present application; Figure 11 This is a structural diagram of a computer device provided in Example 7 of the present application. DETAILED DESCRIPTION

[0013] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0014] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0015] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0016] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0017] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0018] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0019] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0020] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0021] It should be understood that the size of the serial numbers of the steps in the following embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0022] In order to illustrate the technical solution of the present application, specific embodiments are provided below.

[0023] The dental image processing method provided in the first embodiment of the present application can be applied to Figure 1 In an application environment, a client communicates with a server. The dental image processing method can be used on a server, where the client provides the server with a corresponding panoramic dental film, so that the server executes the dental image processing method to analyze the panoramic dental film. Of course, the dental image processing method can also be used on a client, where the server provides the client with a corresponding panoramic dental film, and the client executes the dental image processing method to process the panoramic dental film and obtain a processing result.

[0024] Clients include, but are not limited to, PDAs, desktop computers, laptops, ultra-mobile personal computers (UMPCs), netbooks, cloud computing devices, personal digital assistants (PDAs), and other computer devices. Servers can be implemented as standalone servers or server clusters consisting of multiple servers.

[0025] See also Figure 2 , is a flow chart of a dental image processing method provided in Example 2 of this application, and the dental image processing method is applied to the client. Figure 2 As shown, the dental image processing method may include the following steps: Step S201: Obtain a panoramic dental radiograph to be analyzed, perform feature extraction on the teeth in the panoramic dental radiograph, and determine the root length of each tooth and the corresponding tooth.

[0026] In this embodiment, the panoramic dental film can be obtained by the client sending a corresponding acquisition instruction to the server. Of course, the server can be connected to the corresponding X-ray equipment to obtain the user's oral dental images collected by the X-ray equipment. Figure 3 The figure shows a schematic diagram of tooth feature extraction.

[0027] A neural network model can be used to extract features from the image. For this panoramic dental radiograph, the features are the teeth. Accordingly, the teeth can be extracted by describing their boundaries. Combined with these boundaries, the root length (i.e., the length from the crown to the root) of the corresponding tooth can be determined. The teeth extracted from the feature extraction can be numbered for subsequent calculations and output. This numbering can use internationally accepted tooth position representations (e.g., the FDI system or the ADA system).

[0028] The neural network model can use a model architecture that includes an encoder and a decoder, for example, the YOLOv1-YOLOv5 model.

[0029] Optionally, feature extraction is performed on the teeth in the panoramic dental film to determine the root length of each tooth and the corresponding tooth, including: Obtain a trained YOLOv5 model and use it to extract features from the teeth in the panoramic dental radiographs to determine the root length of each tooth and the corresponding tooth. The training process of the YOLOv5 model is as follows: Acquire a first training image set, the first training image set including at least one panoramic dental film and tooth boundary annotations corresponding to the panoramic dental film; Performing data expansion on the panoramic dental radiographs in the first training image set to obtain an expanded first training image set; Use the expanded first training image set to train the YOLOv5 model to obtain a pre-trained YOLOv5 model; Acquire a second training image set, the second training image set including at least one panoramic dental film and periodontal pocket boundary annotations and bone loss area annotations corresponding to the panoramic dental film; The pre-trained YOLOv5 model is retrained using the second training image set to obtain a trained YOLOv5 model.

[0030] To address the issue of insufficient dental image samples, this implementation employed data augmentation techniques (such as rotation, flipping, and contrast adjustment) to generate more diverse training data. Furthermore, the introduction of new training samples expanded the data source and enhanced the model's ability to recognize diverse dental features. After optimization, the model's overall recognition accuracy reached 80%-90%, significantly improving diagnostic reliability.

[0031] To address the need for identifying key areas of periodontitis, this invention relabeled the periodontitis areas in the training data, ensuring more accurate and comprehensive labeling. By expanding the amount of labeled data and retraining the model, the model's ability to identify key areas (such as periodontal pockets and areas of bone loss) was significantly improved. This high-precision labeling strategy provides higher-quality input for model training, further enhancing the accuracy of lesion analysis.

[0032] Step S202 : segmenting the periodontal bone region in the panoramic dental film to obtain the periodontal bone boundary representing each tooth, and segmenting the enamel region in the panoramic dental film to obtain the enamel boundary representing each tooth.

[0033] In this embodiment, the neural network model is also used for the segmentation of panoramic dental films. Since there are certain differences between regional segmentation and the above-mentioned feature extraction, a model different from the above-mentioned feature extraction can be selected for regional segmentation to achieve better image segmentation effect. Of course, when a model has both feature extraction and segmentation capabilities, the same model can also be used for feature extraction and segmentation. Figure 4 Shown is a schematic diagram of the extraction of periodontal bone boundary and enamel boundary.

[0034] For image segmentation, segmentation area annotations need to be given when training the model, and the segmentation results can be obtained by encoding and decoding.

[0035] Optionally, segmenting the periodontal bone region in the panoramic dental film to obtain a periodontal bone boundary representing each tooth, and segmenting the enamel region in the panoramic dental film to obtain an enamel boundary representing each tooth, including: Obtain a trained U-Net model, and use the trained U-Net model to segment the periodontal bone region in the panoramic dental film to obtain the periodontal bone boundary representing each tooth, and segment the enamel region in the panoramic dental film to obtain the enamel boundary representing each tooth; The training process of the U-Net model is as follows: Acquire a third training image set, wherein the first training image set includes at least one panoramic dental film and periodontal bone boundary annotations and enamel boundary annotations corresponding to the panoramic dental film; performing data expansion on the panoramic dental radiographs in the third training image set to obtain an expanded third training image set; The U-Net model is trained using the expanded third training image set to obtain a trained U-Net model.

[0036] To address the issue of insufficient dental image samples, this implementation employed data augmentation techniques (such as rotation, flipping, and contrast adjustment) to generate more diverse training data. Furthermore, the introduction of new training samples expanded the data source and enhanced the model's ability to recognize diverse dental features. After optimization, the model's overall recognition accuracy reached 80%-90%, significantly improving diagnostic reliability.

[0037] The U-Net model utilizes a convolutional neural network architecture, consisting of multiple convolutional, pooling, and fully connected layers. Input data first undergoes a series of convolution operations. In each layer, the convolution kernel automatically learns local features of the input data. Downsampling in the pooling layer then reduces the data's dimensionality while retaining key information in the image. Finally, multiple fully connected layers further process these features to complete the classification or regression task. Convolutional layers: These layers extract local features from the input data. By using different kernel sizes and stride sizes, these layers can adaptively process different types of input data. Pooling layers: These layers downsample features using max or average pooling, reducing computational effort and enhancing feature invariance. Fully connected layers: These layers process the features extracted by the convolution and pooling layers to produce the final classification or regression output.

[0038] By setting adaptive convolution kernel size and step size, the convolution operation can be dynamically adjusted to adapt to different data characteristics. In traditional convolutional neural networks, the size and step size of the convolution kernel are usually fixed. However, this application uses adaptive convolution kernel settings to automatically optimize parameters according to data changes during training, thereby improving the model's performance on different datasets.

[0039] The pooling layer in this application uses a strategy that combines maximum pooling and average pooling to further extract key information and reduce computational complexity. Through reasonable pooling operations, this application can effectively reduce data redundancy and maintain efficient computing power.

[0040] This application can also introduce the Dropout layer and Batch Normalization technology. The Dropout layer reduces the complexity of the model by randomly discarding a part of the neuron connections, while the Batch Normalization technology ensures the training stability and convergence speed of the network by normalizing the input of each layer, which can prevent overfitting.

[0041] The convolutional neural network proposed in this application is adaptive and can dynamically adjust the parameters of the convolution kernel and network layers based on the characteristics of the input data. By using an adaptive learning rate and a strategy for dynamically adjusting the convolution kernel size, this application can automatically optimize the architecture of the convolutional neural network for different tasks, improving processing efficiency and classification accuracy.

[0042] Step S203: Acquire the target tooth, and determine the bone loss distance between the periodontal bone and the enamel of the target tooth based on the periodontal bone boundary and the enamel boundary of the target tooth.

[0043] In this embodiment, the target tooth may refer to a tooth that meets the preset periodontitis conditions. The specific method for determining the target tooth can refer to the following embodiment. By analyzing the periodontal bone boundary (PBL) and enamel boundary (CEJ) at the target tooth, the distance between the periodontal bone and the enamel can be determined. This distance is used as the bone loss distance (i.e., relative bone loss (RBL)). Figure 5 As shown, it represents the intersection of the upper boundary of the periodontal bone boundary and the upper boundary of the enamel boundary with the central axis in the upper teeth. The distance between the intersections is the bone loss distance.

[0044] Among them, when determining the distance between the periodontal bone and the enamel, the distance between the two nearest boundaries of the periodontal bone boundary and the enamel boundary is selected. For example, the periodontal bone boundary and the enamel boundary are both fitted as straight lines, the distance between the two is detected, and the distance is used as the bone loss distance. For another example, the root position and the middle position of the crown of the target tooth are determined, the root position and the middle position of the crown are connected, and the distance between the connecting line and the intersection of the periodontal bone boundary and the enamel boundary is used as the bone loss distance. For another example, the central axis of the target tooth is determined according to the boundary of the target tooth, and the distance between the central axis and the intersection of the periodontal bone boundary and the enamel boundary is used as the bone loss distance.

[0045] Step S204: determining the periodontitis characterization value of the target tooth based on the bone loss interval and the tooth root length, and determining the staging analysis result of the panoramic dental radiograph based on the periodontitis characterization value.

[0046] In this embodiment, the bone loss interval represents the degree of bone loss, and the tooth root length represents the actual length of the target tooth. Comparing the bone loss interval with the tooth root length can determine the periodontitis marker of the target tooth, thereby determining the periodontitis staging analysis result. For example, if the bone loss interval is larger than the tooth root length, the periodontitis marker is larger and the corresponding periodontitis staging analysis result is larger, i.e., the degree of periodontitis is more severe. If the bone loss interval is smaller than the tooth root length, the periodontitis marker is smaller and the corresponding periodontitis staging analysis result is smaller, i.e., the degree of periodontitis is less severe.

[0047] The embodiment of the present application obtains a panoramic dental film to be analyzed, extracts features of the teeth in the panoramic dental film, determines the root length of each tooth and the corresponding tooth, segments the periodontal bone area in the panoramic dental film, obtains the periodontal bone boundary representing each tooth, segments the enamel area in the panoramic dental film, obtains the enamel boundary representing each tooth, obtains the target tooth, determines the bone loss distance between the periodontal bone and enamel of the target tooth based on the periodontal bone boundary and enamel boundary of the target tooth, determines the periodontitis characterization amount of the target tooth based on the bone loss distance and the root length, and determines the staging analysis result of the panoramic dental film based on the periodontitis characterization amount. By analyzing the tooth length and the periodontal bone and enamel boundaries, it is possible to classify the image features of the teeth in the panoramic dental film, automatically calculate the bone loss, and determine the stage of periodontitis in combination with the tooth length, thereby improving the automation and accuracy of the analysis.

[0048] See also Figure 6 , is a flow chart of a dental image processing method provided in Example 3 of this application. Figure 6 As shown, after extracting features of the teeth in the panoramic dental radiograph and determining the root length of each tooth and the corresponding tooth in step S201, the following steps may be further included: Step S601: extract the periodontal pocket depth of each tooth.

[0049] In step S602 , for any tooth, if the periodontal pocket depth of the tooth is greater than a preset depth value, the tooth is determined to be a target tooth with periodontitis.

[0050] Among them, feature extraction technology is used to extract the periodontal pocket depth of each tooth. Specifically, the periodontal pocket area is identified through a neural network model, and the periodontal pocket depth can be determined by analyzing the periodontal pocket area.

[0051] When the periodontal pocket depth is greater than the preset depth value, the corresponding tooth is determined to be a tooth with periodontitis, that is, a target tooth. When the periodontal pocket depth is not greater than the preset depth value, the corresponding tooth does not have periodontitis and is not a target tooth. In one embodiment, the preset depth value is 2 mm.

[0052] The embodiment of the present application determines whether a tooth has periodontitis by extracting the depth of the periodontal pocket, thereby improving the accuracy of target recognition and effectively improving recognition efficiency.

[0053] See also Figure 7 , is a flow chart of a dental image processing method provided in Example 4 of this application. Figure 7 As shown, after extracting features of the teeth in the panoramic dental radiograph and determining the root length of each tooth and the corresponding tooth in step S202, the following steps may be further included: Step S701: extract the root length of each tooth.

[0054] Step S702: For any tooth, determine the bone loss ratio of the tooth based on the periodontal pocket depth and root length of the tooth.

[0055] Step S703: traverse all teeth to obtain the bone loss ratio of each tooth, and mark the bone loss ratio of each tooth in the panoramic dental film.

[0056] The root length of each tooth is extracted using feature extraction technology. Specifically, the root area is identified through a neural network model, and the root length can be determined by analyzing the root area.

[0057] Comparison of root length and periodontal pocket depth can determine the amount of bone remaining, and thus the proportion of bone loss. The difference is obtained by subtracting the preset length value from the periodontal pocket depth. The root length is compared with the difference to obtain the amount of bone remaining. The proportion of bone loss is obtained by multiplying (1-bone remaining) by 100%.

[0058] After obtaining the bone loss ratio of all teeth, mark the bone loss ratio of each tooth on the panoramic dental film. The embodiment of the present application establishes a mathematical model to calculate the proportion of bone loss in teeth, thereby transforming the traditional subjective diagnosis into a quantifiable objective indicator, allowing doctors to intuitively understand the extent of the lesion, facilitating the planning of subsequent treatment and evaluating the efficacy. The results are more scientific and credible, and at the same time improve the accuracy of disease staging.

[0059] See also Figure 8 , is a flow chart of a dental image processing method provided in Example 5 of this application. Figure 8 As shown, the above step S203 of determining the bone loss distance between the periodontal bone and the enamel of the target tooth according to the periodontal bone boundary and the enamel boundary of the target tooth may include the following steps: Step S801: extract the central axis of the target tooth.

[0060] Step S802: If the target tooth is an upper tooth, determine the first intersection of the upper boundary of the periodontal bone boundary and the central axis, and determine the second intersection of the upper boundary of the enamel boundary and the central axis. The distance between the first intersection and the second intersection is extracted as the bone loss distance between the periodontal bone and enamel of the target tooth.

[0061] Step S803: If the target tooth is a lower tooth, determine the third intersection of the lower boundary of the periodontal bone boundary and the central axis, and determine the fourth intersection of the lower boundary of the enamel boundary and the central axis. The distance between the third intersection and the fourth intersection is extracted as the bone loss distance between the periodontal bone and enamel of the target tooth.

[0062] Among them, there are two types of teeth, upper teeth and lower teeth. Therefore, when identifying the periodontal bone boundary, the periodontal bone boundary of the upper teeth and the lower teeth as a whole can be identified together, while the enamel boundary of the upper teeth and the enamel boundary of the lower teeth cannot be identified together.

[0063] The central axis of the target tooth is extracted. Specifically, the central axis is obtained by analyzing the boundary of the target tooth. If the target tooth is an upper tooth, the first intersection of the central axis and the upper boundary of the periodontal bone boundary, the second intersection of the central axis and the upper boundary of the enamel boundary, and the distance between the first intersection and the second intersection are the bone loss intervals. If the target tooth is a lower tooth, the third intersection of the lower boundary of the periodontal bone boundary and the central axis, the fourth intersection of the lower boundary of the enamel boundary and the central axis, and the distance between the third intersection and the fourth corner point are the bone loss intervals.

[0064] The embodiment of the present application can calculate the bone loss of different teeth, and use the intersection of the central axis and the periodontal bone boundary and the enamel boundary to characterize the bone loss distance, so as to accurately obtain the bone loss distance.

[0065] See also Figure 9, is a flow chart of a dental image processing method provided in Example 6 of this application. Figure 9 As shown, the above step S204 of determining the periodontitis characterization value of the target tooth based on the bone loss interval and the tooth root length, and determining the staging analysis result of the panoramic dental radiograph based on the periodontitis characterization value, may include the following steps: In step S901, the bone loss interval is compared with the tooth root length to obtain the relative bone loss as the periodontitis characterization value of the target tooth. All target teeth are traversed to determine the relative bone loss of all target teeth. The relative bone loss of all target teeth is averaged to obtain the average bone loss value.

[0066] Step S902, comparing the average bone loss value with a preset periodontitis cycle range to determine the periodontitis cycle range in which the average bone loss value lies, and determining that the periodontitis cycle corresponding to the periodontitis cycle range is the staging analysis result of the panoramic dental radiograph.

[0067] In this embodiment, the relative bone loss can be obtained by comparing the bone loss interval with the tooth root length. The relative bone loss is used as a characterization of periodontitis, wherein a higher relative bone loss indicates a more severe periodontitis symptom, and a lower relative bone loss indicates a milder periodontitis symptom.

[0068] The experimental environment used in this application example includes NVIDIA GPU hardware and PyTorch and Flask software. The experimental dataset consists of original panoramic X-rays and their PBL and CEJ boundary labels. The training set, after data augmentation, contains 5,000 images. The experimental steps include image preprocessing, training U-Net and YOLOv5 models, testing model performance, and optimizing parameters. Results analysis: The confusion matrix shows a classification accuracy of 92%, and denoising improves model accuracy by 5%.

[0069] Corresponding to the dental image processing method of the above embodiment, Figure 10 The following is a block diagram of the structure of the dental image processing device provided in Example 7 of the present application. For ease of explanation, only the parts related to the embodiment of the present application are shown.

[0070] See also Figure 10 , the dental image processing device comprises: The tooth feature analysis module 101 is used to obtain a panoramic dental radiograph to be analyzed, extract features of the teeth in the panoramic dental radiograph, and determine the root length of each tooth and the corresponding tooth; The periodontal feature analysis module 102 is used to segment the periodontal bone region in the panoramic dental film to obtain the periodontal bone boundary representing each tooth, and segment the enamel region in the panoramic dental film to obtain the enamel boundary representing each tooth; The bone loss calculation module 103 is used to obtain a target tooth and determine the bone loss distance between the periodontal bone and the enamel of the target tooth based on the periodontal bone boundary and the enamel boundary of the target tooth; The periodontal staging analysis module 104 is used to determine the periodontitis characterization value of the target tooth based on the bone loss interval and the tooth root length, and determine the staging analysis result of the panoramic dental radiograph based on the periodontitis characterization value.

[0071] Optionally, the dental image processing device further includes: The periodontal pocket extraction module is used to extract the characteristics of the teeth in the panoramic dental film, determine the root length of each tooth and the corresponding tooth, and then extract the periodontal pocket depth of each tooth; The target tooth determination module is used to determine, for any tooth, if the periodontal pocket depth of the tooth is greater than a preset depth value, that the tooth is a target tooth with periodontitis.

[0072] Optionally, the dental image processing device further includes: A tooth root length extraction module is used to extract the tooth root length of each tooth after performing feature extraction on the teeth in the panoramic dental film and determining the tooth root length of each tooth and the corresponding tooth; Bone loss calculation module, used to determine the bone loss ratio of any tooth based on the periodontal pocket depth and root length of the tooth; The bone loss marking module is used to traverse all teeth, obtain the bone loss ratio of each tooth, and mark the bone loss ratio of each tooth in the panoramic dental film.

[0073] Optionally, the bone loss calculation module 103 includes: A central axis extraction unit, used to extract the central axis of the target tooth; a first bone loss calculation unit, configured to, if the target tooth is an upper tooth, determine a first intersection point between an upper boundary of the periodontal bone boundary and the central axis, and determine a second intersection point between an upper boundary of the enamel boundary and the central axis, and extract a distance between the first intersection point and the second intersection point as a bone loss distance between the periodontal bone and the enamel of the target tooth; The second bone loss calculation unit is used to determine the third intersection of the lower boundary of the periodontal bone boundary and the central axis, and to determine the fourth intersection of the lower boundary of the enamel boundary and the central axis if the target tooth is a lower tooth, and to extract the distance between the third intersection and the fourth intersection as the bone loss distance between the periodontal bone and the enamel of the target tooth.

[0074] Optionally, the tooth feature analysis module 101 includes: A tooth feature analysis unit is used to obtain a trained YOLOv5 model, use the trained YOLOv5 model to extract features of teeth in a panoramic dental radiograph, and determine the root length of each tooth and the corresponding tooth; The training process of the YOLOv5 model is as follows: Acquire a first training image set, the first training image set including at least one panoramic dental film and tooth boundary annotations corresponding to the panoramic dental film; Performing data expansion on the panoramic dental radiographs in the first training image set to obtain an expanded first training image set; Use the expanded first training image set to train the YOLOv5 model to obtain a pre-trained YOLOv5 model; Acquire a second training image set, the second training image set including at least one panoramic dental film and periodontal pocket boundary annotations and bone loss area annotations corresponding to the panoramic dental film; The pre-trained YOLOv5 model is retrained using the second training image set to obtain a trained YOLOv5 model.

[0075] Optionally, the periodontal feature analysis module 102 includes: a periodontal feature analysis unit, configured to obtain a trained U-Net model, and use the trained U-Net model to segment the periodontal bone region in the panoramic dental film to obtain the periodontal bone boundary representing each tooth, and to segment the enamel region in the panoramic dental film to obtain the enamel boundary representing each tooth; The training process of the U-Net model is as follows: Acquire a third training image set, wherein the first training image set includes at least one panoramic dental film and periodontal bone boundary annotations and enamel boundary annotations corresponding to the panoramic dental film; performing data expansion on the panoramic dental radiographs in the third training image set to obtain an expanded third training image set; The U-Net model is trained using the expanded third training image set to obtain a trained U-Net model.

[0076] Optionally, the periodontal staging analysis module 104 includes: a bone loss analysis unit for comparing the bone loss interval with the tooth root length to obtain a relative bone loss as a periodontitis characterization quantity for the target tooth, traversing all target teeth to determine the relative bone loss of all target teeth, and averaging the relative bone loss of all target teeth to obtain an average bone loss value; The periodontal staging analysis unit is used to compare the average bone loss value with the preset periodontitis cycle range, determine the periodontitis cycle range in which the average bone loss value is located, and determine that the periodontitis cycle corresponding to the periodontitis cycle range is the staging analysis result of the panoramic dental film.

[0077] It should be noted that the information interaction, execution process and other contents between the above modules are based on the same concept as the method embodiment of this application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0078] Figure 11 This is a schematic diagram of the structure of a computer device provided in Example 8 of this application. Figure 11 As shown, the computer device of this embodiment includes: at least one processor ( Figure 11 Only one is shown in the figure), a memory, and a computer program stored in the memory and executable on at least one processor, wherein when the processor executes the computer program, the steps of any of the above-mentioned dental image processing method embodiments are implemented.

[0079] The computer device may include, but is not limited to, a processor and a memory. It will be understood by those skilled in the art that Figure 8 The above is merely an example of a computer device and does not constitute a limitation on the computer device. The computer device may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include a network interface, a display screen, and an input device.

[0080] The processor may be a CPU, other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0081] Memory includes readable storage media, internal memory, and the like. Internal memory can be the internal memory of a computer device, providing an environment for the operation of the operating system and computer-readable instructions stored in the readable storage medium. The readable storage medium can be the computer device's hard drive. In other embodiments, it can also be an external storage device, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, or a flash memory card. Furthermore, memory can include both the computer device's internal storage unit and external storage devices. Memory is used to store the operating system, application programs, boot loaders, data, and other programs, such as the program code of computer programs. Memory can also be used to temporarily store data that has been output or is about to be output.

[0082] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned device can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here. If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the process steps in the above-described method embodiments by instructing the relevant hardware through a computer program. The computer program may be stored in a computer-readable storage medium. When executed by a processor, the computer program implements the steps of the above-described method embodiments. The computer program includes computer program code, which may be in source code form, object code form, executable file, or some intermediate form. Computer-readable media may include at least: any entity or device capable of carrying computer program code, recording media, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunications signals, and software distribution media. Examples include USB flash drives, removable hard drives, magnetic disks, or optical disks. In some jurisdictions, based on legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunications signals.

[0083] The present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed through a computer program product. When the computer program product runs on a computer device, the computer device can implement the steps in the above-mentioned method embodiment when executing it.

[0084] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0085] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0086] In the embodiments provided in this application, it should be understood that the disclosed apparatus / computer equipment and methods can be implemented in other ways. For example, the apparatus / computer equipment embodiments described above are merely schematic. For example, the division of modules or units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of the apparatus or unit, which can be electrical, mechanical or other forms.

[0087] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0088] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A dental image processing method, characterized in that: The dental image processing method comprises: Obtaining a panoramic dental film to be analyzed, extracting features of teeth in the panoramic dental film, and determining the root length of each tooth and the corresponding tooth; Segmenting the periodontal bone region in the panoramic dental film to obtain a periodontal bone boundary representing each tooth, and segmenting the enamel region in the panoramic dental film to obtain an enamel boundary representing each tooth; Acquire a target tooth, and determine a bone loss distance between the periodontal bone and the enamel of the target tooth according to the periodontal bone boundary and the enamel boundary of the target tooth; The periodontitis characterization amount of the target tooth is determined according to the bone loss interval and the tooth root length, and the staging analysis result of the panoramic dental film is determined according to the periodontitis characterization amount.

2. The dental image processing method according to claim 1, characterized in that: After extracting features of the teeth in the panoramic dental film and determining the root length of each tooth and the corresponding tooth, the method further includes: The depth of periodontal pockets of each tooth was extracted; For any tooth, if the periodontal pocket depth of the tooth is greater than a preset depth value, the tooth is determined to be a target tooth with periodontitis.

3. The dental image processing method according to claim 2, characterized in that: After extracting features of the teeth in the panoramic dental film and determining the root length of each tooth and the corresponding tooth, the method further includes: Extract the root length of each tooth; For any tooth, determining the bone loss ratio of the tooth according to the periodontal pocket depth and root length of the tooth; All teeth are traversed to obtain the bone loss ratio of each tooth, and the bone loss ratio of each tooth is marked in the panoramic dental film.

4. The dental image processing method according to claim 1, characterized in that: Determining the bone loss distance between the periodontal bone and the enamel of the target tooth according to the periodontal bone boundary and the enamel boundary of the target tooth comprises: Extracting the central axis of the target tooth; If the target tooth is an upper tooth, a first intersection point between the upper boundary of the periodontal bone boundary and the central axis is determined, and a second intersection point between the upper boundary of the enamel boundary and the central axis is determined, and a distance between the first intersection point and the second intersection point is extracted as a bone loss distance between the periodontal bone and the enamel of the target tooth; If the target tooth is a lower tooth, determine the third intersection of the lower boundary of the periodontal bone boundary and the central axis, and determine the fourth intersection of the lower boundary of the enamel boundary and the central axis, and extract the distance between the third intersection and the fourth intersection as the bone loss distance between the periodontal bone and enamel of the target tooth.

5. The dental image processing method according to claim 1, characterized in that: The step of extracting features of the teeth in the panoramic dental film and determining the root length of each tooth and the corresponding tooth includes: Obtaining a trained YOLOv5 model, and using the trained YOLOv5 model to extract features of teeth in the panoramic dental film, and determining the root length of each tooth and the corresponding tooth; The training process of the YOLOv5 model is as follows: Acquire a first training image set, wherein the first training image set includes at least one panoramic dental film and tooth boundary annotations corresponding to the panoramic dental film; Performing data expansion on the panoramic dental radiographs in the first training image set to obtain an expanded first training image set; Using the expanded first training image set, training the YOLOv5 model to obtain a pre-trained YOLOv5 model; Acquire a second training image set, wherein the second training image set includes at least one panoramic dental film and a periodontal pocket boundary annotation and a bone loss area annotation of the corresponding panoramic dental film; The pre-trained YOLOv5 model is retrained using the second training image set to obtain a trained YOLOv5 model.

6. The dental image processing method according to claim 5, characterized in that: The step of segmenting the periodontal bone region in the panoramic dental film to obtain a periodontal bone boundary representing each tooth, and segmenting the enamel region in the panoramic dental film to obtain an enamel boundary representing each tooth, comprises: Obtaining a trained U-Net model, and using the trained U-Net model to segment the periodontal bone region in the panoramic dental film to obtain a periodontal bone boundary representing each tooth, and segmenting the enamel region in the panoramic dental film to obtain an enamel boundary representing each tooth; The training process of the U-Net model is as follows: Acquire a third training image set, wherein the first training image set includes at least one panoramic dental film and periodontal bone boundary annotation and enamel boundary annotation of the corresponding panoramic dental film; Performing data expansion on the panoramic dental radiographs in the third training image set to obtain an expanded third training image set; The U-Net model is trained using the expanded third training image set to obtain a trained U-Net model.

7. The dental image processing method according to any one of claims 1 to 6, characterized in that: Determining the periodontitis characterization amount of the target tooth according to the bone loss interval and the tooth root length, and determining the staging analysis result of the panoramic dental film according to the periodontitis characterization amount, includes: Comparing the bone loss interval with the tooth root length, obtaining a relative bone loss as a periodontitis characterization amount of the target tooth, traversing all target teeth, determining the relative bone loss of all target teeth, averaging the relative bone loss of all target teeth, and obtaining an average bone loss value; The average bone loss value is compared with a preset periodontitis cycle range to determine the periodontitis cycle range in which the average bone loss value is located, and the periodontitis cycle corresponding to the periodontitis cycle range is determined as the staging analysis result of the panoramic dental film.

8. A dental image processing device, characterized in that: The dental image processing device comprises: A tooth feature analysis module is used to obtain a panoramic dental film to be analyzed, extract features of teeth in the panoramic dental film, and determine the root length of each tooth and the corresponding tooth; A periodontal feature analysis module, for segmenting the periodontal bone region in the panoramic dental film to obtain a periodontal bone boundary representing each tooth, and for segmenting the enamel region in the panoramic dental film to obtain an enamel boundary representing each tooth; A bone loss calculation module, used for acquiring a target tooth, and determining a bone loss distance between the periodontal bone and the enamel of the target tooth according to the periodontal bone boundary and the enamel boundary of the target tooth; The periodontal staging analysis module is used to determine the periodontitis characterization amount of the target tooth according to the bone loss interval and the tooth root length, and determine the staging analysis result of the panoramic dental film according to the periodontitis characterization amount.

9. A computer device, characterized in that: The computer device comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, and the processor implements the dental image processing method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the dental image processing method according to any one of claims 1 to 7 is implemented.

Images