Feature Selection Method for Facial Emotion Recognition Based on Improved Brainstorming Optimization Algorithm

Abstract

The invention discloses a facial emotion recognition feature selection method based on an improved brainstorming optimization algorithm, the steps are as follows: (1) extracting expression features from a facial image by using a histogram of direction gradient HOG to form an initial feature vector; (2) calculating the emotion The difference value between the eigenvector and the neutral emotion eigenvector, find all the features that are different due to emotional changes, construct the difference eigenvector, and pass it to the feature selection module; (3) use the improved brainstorming method in the feature selection module The optimized feature selection method finds a feature subset with the least number of features and the highest classifier accuracy; (4) uses the support vector machine determined by the above feature subset as an emotion classifier to classify new facial images, and then The facial emotion recognition is completed; the invention significantly improves the recognition accuracy of the algorithm; reduces the search space of the brainstorming optimization technology, and obviously improves the processing speed of the facial emotion recognition problem.

Description

technical field [0001] The invention relates to a feature selection method for facial emotion recognition, in particular to a feature selection method for facial emotion recognition based on an improved brainstorming optimization algorithm. Background technique [0002] Facial Expression Recognition (FER) is an indirect recognition of emotions in images from facial action coding systems, or direct recognition of human emotions from facial images. Emotion is one of the internal representative symbols of human beings, which plays an important role in human perception, reasoning, planning, decision-making and social activities. With the rapid development of machine vision, big data and information technology, facial emotion recognition has attracted extensive attention from academia and industry in the fields of intelligent perception, human-computer interaction, fatigue driving detection, emotional robotics and video surveillance. Generally speaking, a typical facial emotion ...

AI Technical Summary

Problems solved by technology

After checking the relevant literature, there is currently no efficient brainstorming optimization technique suitable for facial image emotional feature selection

Images