Facial Expression Recognition via Graph Neural Network Feature Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing facial expression recognition methods suffer from low accuracy due to variations in individual expressions, real-time changes, and the influence of external conditions such as background, illumination, and angle.
Innovation Solution
A facial expression recognition method that extracts a first feature from color information of pixels and a second feature from facial key points, combining these features to improve recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If facial expression recognition is performed using traditional methods, then the recognition process is simple, but the accuracy of recognizing facial expression types is low
Solution Approach 1:
The patent segments the facial expression recognition process into multiple independent feature extraction pathways: color information processing, facial key point detection, and texture analysis. Each pathway extracts specific features independently, which are then fused to improve overall recognition accuracy while maintaining modular complexity
Solution Approach 2:
The patent merges multiple feature types (color features, key point features, texture features) into a fused feature representation. This combination allows the system to leverage complementary information from different feature sources, resolving the contradiction between improved accuracy and increased complexity through systematic integration
2Measurement precision
If only color information is used for feature extraction, then the processing is fast, but the recognition accuracy is insufficient due to variations in individual expressions and external conditions
Solution Approach 1:
The system segments feature extraction into distinct modules: color information processing, facial key point detection, and texture analysis. This segmentation allows each module to specialize in extracting specific types of features while maintaining processing efficiency, with results later fused for improved accuracy
Solution Approach 2:
The patent creates a composite feature representation by combining multiple feature types (color, key points, texture) analogous to composite materials. This fusion leverages the strengths of each feature type to compensate for their individual weaknesses under varying conditions, achieving robust recognition accuracy
Data Source
AI summary
In a facial expression recognition method, facial key points are identified as graph nodes. A facial graph structure for a first image is constructed with edges between pairs of the graph nodes based on relationships between the facial key points corresponding to the graph nodes. A first feature of a facial texture is extracted from color information of pixels in the first image. A second feature of the first image is extracted by processing the facial graph structure using a graph neural network (GNN). The first feature and the second feature are combined, to obtain a fused feature. A first expression type of a face in the first image that corresponds to the fused feature is determined. The first expression type is determined from a plurality of facial expression types.