Abstract

Tele-diagnosis or telephone consultation is more commonly used in current pandemic COVID-19 situation where social distancing is need of hour. Speech based emotion recognition will be very helpful in applications such as diagnosis of patient’s condition in telemedicine, to monitor understanding of learner in online teaching-learning platform, monitoring of call center calls, etc. Due to speech emotion recognition (SER) system these applications can become more natural. Many researchers have proposed different methods of feature extractions, feature optimizations and classification for speaker dependent as well as speaker independent speech emotion recognition. There is an utmost need of robust speech emotion recognition for real time applications. Term robustness is related to person, gender, language and cultural variations. Due to subject dependency in emotion, this domain is very challenging for researchers. In this paper different approaches for developing speech emotion recognition that are speaker independent and language independent are discussed in brief. Very few researchers have focused on both approaches together. This study summarizes challenges and limitations of current methods with suggestions towards developing robust speech emotion recognition system.

Full Text
Published version (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