Abstract
AbstractRecognizing emotions in the elderly and disabled people is an essential part in knowing whether they need help. In this paper, a study is presented for audio visual emotion recognition based on Hidden Markov Model (HMM). In the realm of audio visual emotion recognition, feature extraction of audio visual emotion and HMM training are very important issues. Emotion features of speech and facial image sequences are extracted andthe HTK toolkit is adopted to train the hidden Markov models for audio, visual and audio visualmulti-stream emotion recognition. In general, the recognition rates of audio-visual multi-stream HMMs are slightly higher than the audio only HMM and visual only HMM, and the recognition rates of negative emotions areslightly than positive emotions.Keywordshidden markov modelfeature extractionaudio-visual fusionemotion recognition
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
