Safe Driving In The Facial Expression Recognition Technology

In this paper, the technical of facial expression recognition is studied and used for detecting driver's tired degree. At first, we show the research result in this field.Then, facial expression recognition is to utilize computer to accomplish pretreatment, feature extraction and recognition of expression images. Principal Component Analysis (PCA) is studied and some modifications are discussed to improve the performance of PCA. The methods of facial expression recognition based on pixel-level fusion, feature-level fusion and decision-level fusion are discussed. By integrating the latest research results in facial expression recognition, A method of two-dimensional principal component analysis (2DPCA) is studied for facial expression recognition is proposed. Experimental results show that 2DPCA used for feature extraction is more powerful than classical PCA. A novel method based on Fisher ratio and multi-classifier fusion with fuzzy integral is proposed for feature selection. A novel method of fuzzy densities determined adaptively based on particle swarm optimization (PSO) is discussed. The experimental result is satisfying.