Micro-expression recognition method based on graph deep learning and face three-dimensional reconstruction

Abstract

The invention relates to a micro-expression recognition method based on graph deep learning and face three-dimensional reconstruction, which comprises the following steps of: constructing a graph feature learning module, and performing graph feature analysis to obtain a one-dimensional feature vector; constructing an optical flow feature learning module, and obtaining a one-dimensional feature vector through optical flow feature extraction; constructing a three-dimensional detail reconstruction module to obtain a one-dimensional feature vector; and constructing a multi-stream OGC-FL network model structure, and obtaining a micro-expression recognition and classification result through multi-stream fusion. Compared with a single strategy, the optical flow features are generated through multiple strategies, and the generation strategy most beneficial to a micro-expression recognition task can be screened out; according to the multi-stream OGC-FL network model structure, the consistency of face key point information and dense image information in micro expression recognition is found, the key point sparse space information can judge the rough state of the micro expression through GFL, the dense image information highlights the fine muscle movement of the face, and more detailed information is extracted for MER.

Description

technical field [0001] The technical field of image processing of the present invention, specifically relates to a micro-expression recognition method based on graph deep learning and three-dimensional reconstruction of human face. Background technique [0002] Micro-expression recognition has the potential to discover the true emotions of human beings. Micro-expressions usually only last for a very short period of time, about 1 / 25 to 1 / 3 of a second on the face, and only local muscle movements occur, so it is very difficult to observe and recognize them correctly. In addition, microexpressions, as a spontaneous facial feature, are difficult to fake or suppress. Because micro-expression recognition can reveal a person's true feelings, it can be widely used in many fields that need to accurately identify human emotions, such as national defense and security, criminal investigation of cases, psychopathology, social communication, etc. [0003] In recent years, the potential ...

AI Technical Summary

Problems solved by technology

Some traditional machine learning methods focus on using hand-crafted spatio-temporal features to identify subtle facial muscle movements. They are all used to capture the subtle features of facial micro-expressions. In some cases, these spatio-temporal features can even exceed the recognition level of micro-expressions trained specialists

Images