Low-resolution image facial expression recognition method based on feature reconstruction model

Abstract

The invention discloses a low-resolution image facial expression recognition method based on a feature reconstruction model, and belongs to the field of facial image expression recognition. The method comprises the following steps: constructing training and testing data sets; training a facial expression recognition model of the feature reconstruction model, extracting image expression features by using a feature extraction network with fixed parameters, training the model by using a generative adversarial network mode to obtain an expression feature generator and a feature discriminator, and reconstructing features by using FSRG as an input image to obtain FSR; classifying the feature FSR by a classifier consisting of a full-connection network and a softmax function layer, and re-weighting sample loss by using a probability value of a correct category corresponding to a sample output by the softmax layer; the invention is not sensitive to the resolution of the input image, the recognition accuracy under the low resolution is improved, and the recognition effect on each resolution is more stable.

Description

technical field [0001] The invention belongs to the field of face image expression recognition, in particular to a low-resolution image face expression recognition method based on a feature reconstruction model. Background technique [0002] Facial expressions are one of the most direct and natural signals for humans to express emotions. Facial expression recognition is a hot topic in the research of natural human-computer interaction, computer vision, emotional computing and image processing, and has a wide range of applications in the fields of human-computer interaction, distance education, security, intelligent robot development, medical treatment, animation production, etc. . [0003] In different scenarios, due to changes in equipment and environment and the imaging principle of pinhole cameras, the facial images of people in multi-person photography scenarios have different resolutions of "near large, far small". Compression reduces the quality and resolution of an ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to overcome the shortcomings of large amount of calculation and easy disclosure of privacy in reconstructed face images, and provide a low-resolution image facial expression recognition method based on feature reconstruction model

Images