Three-dimensional tooth model grid segmentation method based on local attention mechanism

Abstract

The invention relates to the field of medical image processing and computer vision, particularly relates to a three-dimensional tooth model grid segmentation method based on a local attention mechanism. The method comprises the following steps: judging and classifying each triangular mesh in a three-dimensional tooth model by adopting a trained three-dimensional tooth model segmentation network, determining whether the area where each triangular grid is located belongs to gingiva or a certain tooth, and accurately finding the complete area where each tooth is located in the three-dimensional tooth model, and therefore the three-dimensional tooth model is accurately segmented. According to the method, the segmentation accuracy of low-feature-recognition-degree regions such as tooth edges and adjacent teeth can be effectively improved.

Description

technical field [0001] The invention relates to the fields of medical image processing and computer vision, in particular to a three-dimensional tooth model grid segmentation method based on a local attention mechanism. Background technique [0002] With the development of digital technology, digital stomatology technology is rapidly changing the traditional diagnostic mode in the field of stomatology, in which computer-aided treatment systems are widely used in the field of stomatology, and the accurate segmentation of teeth from the patient's digital three-dimensional tooth model is a computer The important basis of the auxiliary treatment system, its segmentation results can assist the doctor to move and rearrange the patient's teeth to simulate the treatment effect after orthodontics. The development of follow-up treatment plan provides important reference information. Different from ordinary 2D images, 3D tooth models are essentially unstructured data composed of 3D po...

AI Technical Summary

Problems solved by technology

[0003] (1) The shape of the tooth model of different people varies greatly, so the traditional segmentation method based on geometric features is often less robust;

[0004] (2) The patient's teeth are often crowded and misaligned, resulting in unclear tooth boundaries;

[0005] (3) Some patients still have missing teeth, and the noise generated during the scanning process brings additional challenges to the segmentation task

Although these traditional methods are relatively intuitive, they rely heavily on prior knowledge and require a certain amount of manual interaction. Therefore, fully automatic segmentation cannot be achieved and the segmentation performance is easily affected by subjective factors.

Some 3D tooth model segmentation methods based on deep learning are prone to over-segmentation or under-segmentation in tooth edges and adjacent tooth regions because they cannot effectively extract local detailed semantic information.

Images