Facial expression recognition method based on depth measurement fusion network

Abstract

The invention discloses a facial expression recognition method based on a depth measurement fusion network, and the method comprises the following steps: firstly carrying out the preprocessing of an input image; secondly, constructing a depth measurement fusion network combined with various expression feature representations, and finally, training the depth measurement fusion network by utilizingthe training sample; and during testing, inputting a to-be-detected facial expression image into the trained depth measurement fusion network to obtain classification categories of expressions. The method has the following beneficial effects that different expression distributions are learned by using multiple pieces of threshold information, so that the learned features are more robust and distinguishable; a symmetric triple loss function is used, the problems of incomplete judgment and anchor point selection sensitivity can be avoided, convergence of the deep metric fusion network is accelerated, and the learning ability is improved; the accuracy of facial expression recognition can be improved.

Description

technical field [0001] The invention relates to the technical field of facial expression recognition, in particular to a facial expression recognition method based on a deep metric fusion network that can improve the accuracy of facial expression recognition. Background technique [0002] Facial expression is one of the most natural and common signals for humans to convey emotional states. Expression recognition has broad application prospects in business, security, medicine and other fields. The ability to quickly and accurately recognize facial expressions is of great significance to its research and application. Traditional machine learning methods need to manually extract features and the accuracy is difficult to guarantee. In recent years, convolutional neural networks have been widely used in expression recognition due to their good self-learning and generalization capabilities, but expression recognition is still a challenge due to changes in pose, lighting, and indiv...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the difficulty of facial expression classification based on image features in the prior art and the difficult convergence of metric learning, which will lead to insufficient learning ability of the network and low recognition accuracy, the present invention provides a method that can improve the accuracy of facial expression recognition. Efficient facial expression recognition method based on deep metric fusion network

Images