Three-dimensional CBCT tooth image segmentation method based on feature transformation

Abstract

The invention belongs to the technical field of computer vision, and particularly relates to a three-dimensional CBCT tooth image segmentation method based on feature transformation, and the method comprises the steps of obtaining CBCT image data in real time, and carrying out the preprocessing of the data; inputting the preprocessed CBCT image data into the trained CBCT image tooth segmentation model for segmentation processing; evaluating and analyzing the segmentation result, wherein the CBCT image tooth segmentation model is an improved 3D convolutional neural network, and the improved 3D convolutional neural network comprises an encoder, a spatial transformation module STM, a category transformation module CTM, a feature fusion module, a decoder and an output layer. According to the invention, a 3D convolutional neural network model combining a spatial feature transformation module and a category feature transformation module is adopted, spatial global information and category global information are combined, the segmentation effect is effectively improved, and the classification result is improved.

Description

technical field [0001] The invention belongs to the technical fields of oral clinical medicine and computer vision, and in particular relates to a method for segmenting three-dimensional CBCT tooth images based on feature transformation. Background technique [0002] In the field of stomatology, virtual orthodontic systems developed by combining technologies such as computer graphics, digital media, and graphic image processing are assisting dentists in diagnosis and treatment. CBCT images have the characteristics of clear imaging, high resolution and low radiation dose. The 3D tooth model generated after tooth segmentation from 3D CBCT can be used to analyze the relationship between adjacent teeth and guide tooth implantation; it can also be used for mechanical operations, such as measuring the ratio of root and crown for orthodontic treatment. Three-dimensional tooth model has important guiding significance and reference value for scheme selection and implant treatment. ...

AI Technical Summary

Problems solved by technology

[0004] At present, many methods have been proposed to realize the segmentation of teeth in CBCT oral images, but these methods have not been able to solve the following problems: 1) CBCT root apices have large differences in shape, and it is very difficult to rely on local features to segment them. Misclassification and missed classification; 2) Different tooth categories may have large differences in the same category and small differences between categories, which is likely to cause confusion between categories

These problems lead to a great difference between the tooth segmentation results and the precise standard of the actual image tooth boundary, so a new method is urgently needed to segment the 3D CBCT tooth image more accurately

Images