Face emotion recognition method based on deep neural network

Abstract

The invention relates to the technical field of face recognition, in particular to a face emotion recognition method based on a deep neural network. The method comprises the following steps of automatically downloading free videos from the Internet by utilizing a crawler tool; automatically exporting pictures containing face information from the downloaded video by using a face detection tool, and sequentially outputting frames according to a time axis; marking all faces in the exported video by using bounding box coordinates; outputting an image in the bounding box as a training material; wherein a large amount of network data is adopted as training data, collection is convenient, the training data are free data, pictures containing face information are automatically captured through a face detection tool, the pictures are labeled and output, and finally deep learning and training are performed through a DenseXception network model based on DenseNet and Xception to obtain a final model.

Description

technical field [0001] The invention relates to the technical field of face recognition, in particular to a method for face emotion recognition based on a deep neural network. Background technique [0002] Emotion recognition originally refers to the individual's recognition of other people's emotions. Now it mostly refers to AI's automatic identification of the individual's emotional state by obtaining the individual's physiological or non-physiological signals, which is an important part of emotional computing. The content of emotion recognition research includes facial expression, voice, heart rate, behavior, text and physiological signal recognition, etc., through which the user's emotional state can be judged. [0003] This article only describes the case of emotion recognition through facial expressions. [0004] Existing technologies usually use deep learning to realize emotion recognition. For example, a Chinese patent discloses an emotion recognition method and sys...

AI Technical Summary

Problems solved by technology

[0005] The disadvantage of the above-mentioned technical solution is that all the objects of learning and judgment are the employees who clock in for work. Based on the experiment, we know that the employees who clock in for work have a high probability of showing a poker face (not happy or sad, numb, etc.), and the model trained in this way is obviously No real emotion recognition