Facial action unit strength estimation-based expression analysis method

Abstract

The present invention provides a facial action unit strength estimation-based expression analysis method. The method mainly comprises the content of structured depth condition random field, one-element potential, paired potential, learning and deduction, and comprises the process that the structured depth condition random field has two settings, in the first setting, an input facial image is given, a pre-defined convolutional neural network layer is applied to the (normalization) input image to generate a characteristic graph, and the second setting is to use a data enhancement learning method to utilize the information coming from a plurality of data sets; the one-element node potential is defined in a structural depth condition random field, when the random variables are discrete, the joint distribution can be constructed for the discrete variables, and an objective function of the learned and generated structural depth condition random field is enhanced from the plurality of data sets. The expression analysis method of the present invention utilizes the deep structured learning, can process the high-dimension input characteristics, and enables the image characteristics to be improved substantially, the performances to be improved remarkably and the facial action unit strength estimation accuracy to be improved.

Description

technical field [0001] The invention relates to the field of expression analysis, in particular to an expression analysis method based on facial action unit strength estimation. Background technique [0002] The recognition and analysis of facial expressions is an important research direction in the field of artificial intelligence. It not only has universal significance in social life, but also plays an important role in computer emotional computing. It automatically recognizes peoples facial expressions, and then analyzes peoples emotions and psychological activities. It can be applied in the field of security. In public places, such as airports, subway stations, etc., it can automatically analyze peoples emotions through monitoring equipment such as cameras installed. Expressions and movements, through these analyzes to further judge the characters psychology, so as to judge suspicious characters or even terrorists, and then prevent their criminal behavior; it can also he...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problem of low recognition accuracy, the object of the present invention is to provide an expression analysis method based on facial action unit strength estimation. The structured depth conditional random field includes two settings. In the first setting, given the input facial image , which applies a predefined convolutional neural network layer to (normalize) the input image to generate a feature map, and the second setting uses data augmentation learning methods in order to leverage information from multiple datasets; The unary node potential is defined in the airport. When the random variable is discrete, the joint distribution can be constructed for the discrete variable. The objective function of the deep structured conditional random field generated by reinforcement learning from multiple data sets

Images