A small sample tooth detection method and storage medium based on prototype segmentation network

Abstract

The invention is a small-sample tooth detection method and storage medium based on a prototype segmentation network. The network inputs a simulated image and a real image, and the two sample input networks will respectively obtain virtual image features and real image features after CNN feature extraction. The simulated virtual image features are combined with the Mask label, and then feature clustering is performed by the NPR module. The real image features are divided into two branches, one of which is guided by the characteristics of the clustering feature to obtain the preliminary segmentation result, and the other branch is used for the target detection network. The final detected result is fused with the preliminary segmentation result, and the segmentation and detection results are mutually promoted to obtain an accurate tooth segmentation result. The invention effectively solves the problem of tooth segmentation with small samples, and utilizes the respective advantages of target detection and semantic segmentation to promote and learn from each other, solves the problems of wrong tooth segmentation types and inaccurate tooth boundaries, and can save a lot of time and labor costs .

Description

technical field [0001] The invention relates to the technical field of oral medical imaging, in particular to a small sample tooth detection method and a storage medium based on a prototype segmentation network. Background technique [0002] With the development and progress of society, people's living standards have been greatly improved, and their lifestyles have become more and more diversified. This has also produced many bad living habits and caused oral problems. People pay more and more attention to the condition of the oral cavity. According to the oral image, the overall structure of the oral cavity can be observed, and then the case data can be recorded. However, in the existing tooth detection solutions, the problems of wrong tooth segmentation and inaccurate tooth boundaries are particularly prominent, and a large number of samples are often required for model training, which cannot be used in Tooth detection is performed with a small number of dental data and an...

AI Technical Summary

Problems solved by technology

[0006] The purpose of this application is to solve and overcome the deficiencies of existing technologies and applications, to provide a small-sample tooth detection method and storage medium based on the prototype segmentation network, which effectively solves the problem of small-sample tooth segmentation, while using target detection and semantic The respective advantages of segmentation, mutual promotion and mutual learning, solve the problems of wrong tooth segmentation types and inaccurate tooth boundaries, and can save a lot of time and labor costs

Images