期刊名称：International Journal of Electronics and Computer Science Engineering
电子版ISSN：2277-1956
出版年度：2012
卷号：1
期号：3
页码：1791-1798
出版社：Buldanshahr : IJECSE
摘要：The most important and integral part of an effective human-computer interaction system lies in machine recognition of human affective state. Existing research in this domain focus on addressing problems related to utilization of either audio or visual signals. This paper proposes an organized loom to perceive and recognize human affective state from audiovisual information. Audio signals of affective speech are signified by employing Mel-frequency Cepstral Coefficient, the extracted prosodic, and formant frequency features. To depict visual information Gabor wavelet features are employed. A stepwise method based on Mahalanobis distance is effectively employed in order to perform feature selection. Emotions could be classified based on audiovisual features. Performance Analysis among different classification algorithms could be carried out and finally this novel method could be justified to outperform existing techniques
