Shivaji Chaudhari, Ramesh Kagalkar
Abstract: Accurate gender classification is mostly convenient in case of speech and speaker recognition and also in speech emotion classification; since a superior performance has been stated when separate acoustic models are employed for males and females. Gender classification is also specious into face recognition, particular video summarization, human or robot interaction (HCI), etc. In various criminal cases, an evidence either in the form of as phone conversations or in the form of as tape recordings. Thus, act of law enforcement agencies have been concerned which help the identification of a criminal about accurate approaches to profile dissimilar characteristics of a speaker from recorded patterns of voice. The importance of automatically recognizing expressed emotions from human speech has grown with the increasing role of spoken language interfaces in human-computer interaction (HCI) applications. This explores the detection of domain-specific emotions using language and discourse information in conjunction with acoustic correlates of emotion in speech signals. The main motivation is on a case study of detecting negative and non-negative emotions using spoken language data obtained from a call center application. Many previous surveys in emotion identification have used only the acoustic information contained in speech.
Keywords: age estimation, gender detection, Gaussian Mixture Model GMM, Hidden Markov Model HMM, Mel Frequency Cepstral Coefficients MFFCs, dimension reduction, speaker emotion recognition, weighted supervised non-negative matrix factorization