Image semantic description method and system based on multi-feature extraction

Abstract

The invention relates to an image semantic description method and system based on multi-feature extraction, belongs to the technical field of image feature extraction, and solves the problems that inthe prior art, extracted image features are single, and a long-short-term memory network can only capture one-way time sequence information. The method comprises the following steps: inputting an image to be semantically described into a global feature extraction model to obtain a global feature vector of the image; extracting an attribute feature vector of the image; simultaneously inputting theglobal feature vector and the attribute feature vector into a bidirectional long-short term memory network to obtain a forward joint loss function and a backward joint loss function; and accumulatingand summing the forward joint loss function and the backward joint loss function, and when the sum is minimum, obtaining semantic description which is optimally matched with the image. The extractionof multiple image features is realized, and the accuracy of semantic description is improved.

Description

technical field [0001] The invention relates to the technical field of image feature extraction, in particular to an image semantic description method and system based on multi-feature extraction. Background technique [0002] Image semantic description has always been one of the most important research directions in the field of artificial intelligence, and it is an advanced task for image understanding. At present, image semantic description methods based on deep neural networks have made major breakthroughs in this field, especially the semantic description generation model that combines convolutional neural networks and recurrent neural networks. [0003] Mao et al. creatively combined convolutional neural network and recurrent neural network to solve problems such as image description and sentence retrieval. After that, Kiros et al. took the lead in introducing the encoding-decoding framework into the research of image semantic description. They utilize deep convoluti...

AI Technical Summary

Problems solved by technology

[0005] In view of the above analysis, the embodiment of the present invention aims to provide a method and system for image semantic description based on multi-feature extraction to solve the problem that the image features extracted in the prior art are single and LSTM can only capture one-way time series information, resulting in semantic description low precision problem

Images