Image semantic understanding analysis method based on global interaction

Abstract

The invention constructs a global interactive image semantic analysis method based on machine vision, is applied to the generation of an image semantic title, and comprises the following specific steps: 1) selecting a target image feature extraction model, and carrying out feature extraction and coding on image data; 2) establishing a global interactive bidirectional recurrent neural network to analyze image features; 3) carrying out standard regularization processing on the extracted image feature information, and sending the image feature information into a semantic analysis model in real time in a global information mode to carry out model training; and 4) performing semantic analysis on a new target image through the trained model. The image semantics generated by the image semantics understanding model and algorithm established by the invention has the characteristics of strong logicality and rich semantics, has the advantages of high model convergence speed, high semantic analysis precision and the like, and is successfully applied to the fields of content-based image retrieval, medical image analysis, auxiliary blind guiding, early education of children and the like.

Description

technical field [0001] The invention relates to an analysis method for constructing an image semantic understanding model based on global interaction, and is applied to the fields of content-based image retrieval, medical image analysis, auxiliary guidance for the blind, early childhood education and the like. That is, given a picture, generating a text description of the picture belongs to the technical fields of target detection and semantic analysis algorithms. Background technique [0002] Image semantic understanding is based on image recognition and integrates interdisciplinary research in computer science, psychology, and linguistics, and has also made important contributions to the study of cross-modal interactions between images and texts. Image semantic understanding technology wants to rationally or perceptually understand the target image as a whole and generate a natural language description that conforms to human habits. It is a very complex and challenging ta...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problems of the representative model of encoding and decoding structure - the baseline model (Neural Image Caption Generator, NIC), the model convergence speed is slow, the generated semantic analysis logic is poor, etc., the present invention proposes that the global image information The semantic analysis model is sent in real time to guide the semantic generation, and the bidirectional cyclic neural network model is used to analyze the semantics of the image, so as to obtain a more logical and precise image semantic description

Images