Automatic Speaker Age Estimation and Gender Dependent Emotion Recognition

Abstract

Gender-dependent age, emotions (stress and feeling) are speaker qualities being examined in voice-based speaker voice processing system, these qualities or characteristics play important role in the Human and Computer Interaction (HCI). Grouping speaker attributes is an important task in the fields of Voice Processing, Sspeech Synthesis, Forensics, Language Learning, Assessment, furthermore Speaker Identification to increase the performance of voice processing system, also enhance the emotion identification depend on two-stage recognizer that identify the gender of speaker male or female And then recognize the emotions. Noise elimination technique eliminate the noisy sound from audio clip. Mel-Frequency Cepstral Coefficients (MFCCs) is a feature extraction technique broadly utilized as an important part of Automatic voice processing for unique feature extraction. The system contains the Gaussian mixture model (GMM) supervectors as features for a support vector machine (SVM) for the large data classification into different group based on the margin between the two different classes. Principal component analysis (PCA)is used to reduce the large dimension size of feature vector to improve the system performance and accuracy in HCI.