作者: Xiaoyi Feng , Abdenour Hadid , Matti Pietikäinen
DOI: 10.1007/978-3-540-30126-4_81
关键词: Sadness 、 Local binary patterns 、 Expression (mathematics) 、 Facial expression 、 Facial expression recognition 、 Feature vector 、 Pattern recognition 、 Sample (graphics) 、 Artificial intelligence 、 Computer science 、 Classification scheme
摘要: In this paper, a coarse-to-fine classification scheme is used to recognize facial expressions (angry, disgust, fear, happiness, neutral, sadness and surprise) of novel expressers from static images. the coarse stage, seven-class problem reduced two-class one as follows: First, seven model vectors are produced, corresponding basic expressions. Then, distances each vector feature testing sample calculated. Finally, two expression classes selected sample’s candidates (candidate pair). fine K-nearest neighbor classifier fulfils final classification. Experimental results on JAFFE database demonstrate an average recognition rate 77% for expressers, which outperforms reported same database.