Research on Disambiguation Attributes Based on Semantic Hierarchical Model

Abstract

Feature extraction concepts and statistical classifiers began to be used in computer-aided detection and diagnosis, and many different feature learning techniques such as principal component analysis, image block clustering and dictionary methods are very popular. However, in the case of less labeled data, the application and effect of deep learning are restricted. The research content of this paper is based on items such as nodule multi-semantic and fine-grained classification and rating research of WGAN synthetic oversampling learning, multi-semantic assisted disambiguation research based on multi-level association transfer learning, and multi-label incremental learning research for small sample semantic hierarchical structure. Moreover, based on the synthesis of various semantic samples, multi-semantic association transfer, and multi-label incremental learning methods, a set of effective methods and application mechanisms for intelligent auxiliary semantic hierarchical structure disambiguation are formed by describing the problem definition, proposing research methods, and discussing the experimental effects of the proposed methods in detail.