Research on key problems of image understanding based on semantic information
基于语义信息的图像理解关键问题研究

This paper presented two key problems to shorten "semantic gap" distance between low-level visual features and high-level semantic features.First,according to the abstract degree of semantic information,set up a kind of image hierarchical model,and analyzed the characteristics and limitations of four semantic extraction methods in detail.Then,compared some representative similarity measures,and illustrated application research status in this field.Finally,based on the framework of image semantic understandi...