CN-122023666-A - Dental three-dimensional digital model coordinate system analysis method, device, equipment and medium

CN122023666ACN 122023666 ACN122023666 ACN 122023666ACN-122023666-A

Abstract

The invention relates to the technical field of image processing and discloses a method, a device, equipment and a medium for analyzing a dental three-dimensional digital model coordinate system, which comprise the steps of obtaining a dental three-dimensional digital model of a target user, calculating an occlusion plane of a dental region in the dental three-dimensional digital model, and extracting a normal vector of the occlusion plane; mapping a single-jaw model in a dental three-dimensional digital model into a dental projection depth map according to an occlusion plane and a normal vector, extracting multi-scale depth features of the dental projection depth map, carrying out pooling operation on the multi-scale depth features to obtain dental types, carrying out up-sampling fusion on the multi-scale depth features to generate a two-dimensional key point heat map, calculating a dental region direction vector of a target tooth in the dental three-dimensional digital model according to the dental region direction vector, the normal vector and the dental type, and determining a dental coordinate system of the dental three-dimensional digital model according to the dental region direction vector, the normal vector and the dental type. The accuracy of the dental three-dimensional digital model coordinate system analysis is improved.

Inventors

XIE SHICHEN
She yue

Assignees

成都齿灵科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260211

Claims (10)

1. A method for analyzing a dental three-dimensional digital model coordinate system, the method comprising: Acquiring a dental three-dimensional digital model of a target user, calculating an occlusion plane of a dental region in the dental three-dimensional digital model, and extracting a normal vector of the occlusion plane; mapping a single jaw model in the dental three-dimensional digital model into a dental projection depth map according to the occlusion plane and the normal vector; Extracting multi-scale depth features of the dental projection depth map by using a pre-trained neural network model, and carrying out pooling operation on the multi-scale depth features to obtain dental types of target teeth; Upsampling and fusing the multi-scale depth features to generate a two-dimensional key point heat map of the target teeth in the dental projection depth map; calculating a dental region direction vector of a target tooth in the dental three-dimensional digital model according to the dental type and the two-dimensional key point heat map; and determining a dental coordinate system of the dental three-dimensional digital model according to the dental region direction vector, the normal vector and the dental type.
2. The method of claim 1, wherein the calculating the occlusal plane of the dental region in the dental three-dimensional digital model comprises: Calculating the space Euclidean distance between each vertex in the upper jaw and each vertex in the lower jaw of the dental three-dimensional digital model; when the spatial Euclidean distance is smaller than a preset distance threshold, the vertexes corresponding to the spatial Euclidean distance are aggregated into an occlusion region point cluster; Calculating the occlusion plane parameters of a dental region in the dental three-dimensional digital model according to the occlusion region point cluster; And determining the occlusion plane of the dental region in the dental three-dimensional digital model according to the occlusion plane parameters.
3. The dental three-dimensional digital model coordinate system analysis method according to claim 1, wherein the mapping of a single jaw model of the dental three-dimensional digital model to a dental projection depth map according to the occlusal plane and the normal vector comprises: calculating a centroid point in the dental three-dimensional digital model, and calculating a transmitting direction and a ray starting point of single-jaw model mapping according to the normal vector and the centroid point; Determining an emission plane mapped by the single-jaw model according to the occlusion plane and the ray starting point; generating parallel ray clusters of a single jaw model in the dental three-dimensional digital model according to the uniform grid distribution in the emission plane based on the emission direction, and recording the intersection point of each ray in the parallel ray clusters and the surface of the single jaw model; Calculating Euclidean distance between the intersection point and the ray starting point, and quantizing the Euclidean distance into a depth value of a pixel at a position corresponding to the intersection point; And combining the depth values of all the pixel points into a dental projection depth map.
4. The method of claim 1, wherein extracting multi-scale depth features of the dental projection depth map using a pre-trained neural network model comprises: Performing convolution operation on the dental projection depth map by using a convolution layer in a pre-trained neural network model to obtain a preliminary image feature of the dental projection depth map; performing nonlinear transformation on the preliminary image features to obtain updated image features of the dental projection depth map; And performing downsampling operation on the updated image features to obtain multi-scale depth features of the dental projection depth map.
5. The method for analyzing a dental three-dimensional digital model coordinate system according to claim 1, wherein the pooling operation is performed on the multi-scale depth features to obtain a dental type of a target tooth, and the method comprises: Extracting an intermediate layer feature map in the multi-scale depth feature, and performing global average pooling operation on the intermediate layer feature map to obtain a global feature vector; Performing nonlinear transformation on the global feature vector to obtain probability distribution of dental type; determining the jaw type of the target tooth according to the probability distribution.
6. The method for analyzing a dental three-dimensional digital model coordinate system according to claim 1, wherein the up-sampling fusion is performed on the multi-scale depth features to generate a two-dimensional key point heat map of the target teeth in the dental projection depth map, which comprises: upsampling the multi-scale depth features to obtain an upsampled feature map; adding the up-sampling feature map and the same-scale resolution features of a network shallow layer in the neural network model element by element through jump connection to obtain a resolution feature map; and carrying out convolution calculation on the resolution characteristic map to obtain a two-dimensional key point heat map of the target teeth in the dental projection depth map.
7. The method for analyzing a dental three-dimensional digital model coordinate system according to claim 1, wherein the calculating a dental region direction vector of a target tooth in the dental three-dimensional digital model according to the dental type and the two-dimensional keypoint heat map comprises: Extracting coordinates of key points in the two-dimensional key point heat map, and carrying out vector calculation on the coordinates to obtain a preliminary tooth area direction vector of the target tooth; if the dental state in the dental type is a preset first type, taking the preliminary dental region direction vector as a dental region direction vector of the dental state; If the dental state in the dental state is a preset second type, intercepting a half-side curve section corresponding to a preset standard dental arch curve according to the missing side in the dental state; The point cluster of the occlusion area is projected and mapped to a two-dimensional plane according to the emission direction to obtain a point cluster, the point cluster and the half-side curve segment are aligned and fitted through a least square method, and the rotation deflection angle when the point cluster and the half-side curve segment are aligned is calculated; and carrying out rotation correction on the preliminary dental region direction vector according to the rotation deflection angle to obtain the dental region direction vector in the dental state.
8. A dental three-dimensional digital model coordinate system analysis device, the device comprising: The occlusion plane and normal vector extraction module is used for acquiring a dental three-dimensional digital model of a target user, calculating an occlusion plane of a dental region in the dental three-dimensional digital model, and extracting a normal vector of the occlusion plane; The dental projection depth map mapping module is used for mapping a single-jaw model in the dental three-dimensional digital model into a dental projection depth map according to the occlusion plane and the normal vector; The dental type determining module is used for extracting multi-scale depth features of the dental projection depth map by utilizing a pre-trained neural network model, and carrying out pooling operation on the multi-scale depth features to obtain dental types of target teeth; The two-dimensional key point heat map up-sampling module is used for up-sampling and fusing the multi-scale depth features to generate a two-dimensional key point heat map of the target teeth in the dental projection depth map; The dental region direction vector calculation module is used for calculating the dental region direction vector of the target tooth in the dental three-dimensional digital model according to the dental type and the two-dimensional key point heat map; and the dental coordinate system determining module is used for determining a dental coordinate system of the dental three-dimensional digital model according to the dental region direction vector, the normal vector and the dental type.
9. An electronic device, the electronic device comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the dental three-dimensional digital model coordinate system analysis method according to any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the dental three-dimensional digital model coordinate system analysis method according to any one of claims 1 to 7.

Description

Dental three-dimensional digital model coordinate system analysis method, device, equipment and medium Technical Field The present invention relates to the field of image processing technologies, and in particular, to a method, an apparatus, a device, and a medium for analyzing a dental three-dimensional digital model coordinate system. Background The dental three-dimensional digital model is a digital geometric model which is formed by performing optical scanning on teeth, jawbone and peripheral soft tissues of a patient through high-precision oral cavity scanning equipment, acquiring massive point cloud data of the surface morphology of the dental three-dimensional digital model, and performing triangular gridding reconstruction to form the dental three-dimensional digital model which can accurately reflect the three-dimensional morphology and the spatial relationship of an oral anatomy structure. In the field of computer-aided dental diagnosis and treatment, a dental three-dimensional digital model is a core data base for realizing personalized diagnosis and treatment and accurate restoration, however, when high-precision oral cavity scanning devices of different manufacturers output a three-dimensional grid model, an initial space coordinate system which is independent of each other and is set at will is generally adopted, so that model data lack of uniform pose and orientation references. The multi-source heterogeneous coordinate system difference ensures that the subsequent automatic processing flows of AI auxiliary diagnosis, occlusion analysis, prosthesis design, orthodontic scheme planning and the like cannot be directly carried out, and the automatic processing flows are seriously dependent on manual correction operations such as complicated manual rotation, translation and the like by doctors. The manual intervention mode is low in efficiency, unstable in precision due to the introduction of subjective errors, and more difficult to meet the requirements of clinical large-scale data processing, and becomes a key technical bottleneck for restricting the improvement of diagnosis and treatment efficiency and standardization level. Therefore, in the prior art, because the high-precision oral cavity scanning devices of different manufacturers generally adopt the initial space coordinate systems which are independent of each other and are set at will when outputting the three-dimensional grid model, the model data lack uniform pose and orientation reference, and the precision of the dental three-dimensional digital model coordinate system is lower when analyzing. Disclosure of Invention The invention provides a method, a device, equipment and a medium for analyzing a dental three-dimensional digital model coordinate system, and mainly aims to solve the problem of lower accuracy in the process of analyzing the dental three-dimensional digital model coordinate system. In order to achieve the above object, the present invention provides a dental three-dimensional digital model coordinate system analysis method, comprising: Acquiring a dental three-dimensional digital model of a target user, calculating an occlusion plane of a dental region in the dental three-dimensional digital model, and extracting a normal vector of the occlusion plane; mapping a single jaw model in the dental three-dimensional digital model into a dental projection depth map according to the occlusion plane and the normal vector; Extracting multi-scale depth features of the dental projection depth map by using a pre-trained neural network model, and carrying out pooling operation on the multi-scale depth features to obtain dental types of target teeth; Upsampling and fusing the multi-scale depth features to generate a two-dimensional key point heat map of the target teeth in the dental projection depth map; calculating a dental region direction vector of a target tooth in the dental three-dimensional digital model according to the dental type and the two-dimensional key point heat map; and determining a dental coordinate system of the dental three-dimensional digital model according to the dental region direction vector, the normal vector and the dental type. Optionally, the calculating the occlusal plane of the dental jaw region in the dental jaw three-dimensional digital model includes: Calculating the space Euclidean distance between each vertex in the upper jaw and each vertex in the lower jaw of the dental three-dimensional digital model; when the spatial Euclidean distance is smaller than a preset distance threshold, the vertexes corresponding to the spatial Euclidean distance are aggregated into an occlusion region point cluster; Calculating the occlusion plane parameters of a dental region in the dental three-dimensional digital model according to the occlusion region point cluster; And determining the occlusion plane of the dental region in the dental three-dimensional digital model according to the occlusion plane parameters.