Depression recognition system based on micro-expression analysis

Abstract

The invention discloses a depression recognition system based on micro-expression analysis. The invention belongs to the field of computer vision. The method comprises the following specific steps: 1,training a deep multi-task identification network; 2, dividing an important local area of the human face, and removing areas irrelevant to micro-expressions; 3, training an adaptive double-flow neural network, and positioning a start frame, an Apex frame and an end frame of the micro-expression motion; and 4, judging whether the person suffers from depression or not according to the analysis of the micro-expression under different backgrounds. According to the invention, the image is preprocessed on the basis of the deep multi-task neural network, so that important local areas of the face aredivided, the recognition speed of the double-flow neural network is increased, and the requirement for real-time performance is met; important frame picture features are extracted through a BLSTM-CNNneural network based on an attention mechanism, and double-flow features extracted by the double-flow neural network are adaptively fused to improve the positioning of micro-expression motion frames,so that the speed and accuracy of micro-expression recognition are improved.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a depression recognition system based on micro-expression analysis. Background technique [0002] At present, depression is the second largest human disease after cardiovascular disease. According to the official website of the World Health Organization, as of 2017, the number of patients with depression worldwide has exceeded 300 million, and about 800,000 people suffer from depression every year. Suicide. At the same time, the incidence of depression has begun to show a trend of younger age (university, even primary and middle school students). However, the medical prevention and treatment of depression in our country is still in a situation of low recognition rate. Hospitals above prefecture-level cities have a recognition rate of less than 20%, and only less than 10% of patients have received relevant drug treatment. Recognition is critical to medical prevention effo...

AI Technical Summary

Problems solved by technology

[0004] In view of the above problems, the present invention provides a depression recognition system based on micro-expression analysis to solve the problem that the relative displacement between frames is very small due to factors such as weak micro-expression intensity and short duration in the prior art, and it is difficult to detect micro-expression movements. The problem of positioning and the speed of micro-expression recognition

Images