Emotion recognition from speech using MFCC and DWT for security system

In recent years the emotion recognition from speech is area of more interest in human computer interaction. There are many different researchers which worked on emotion recognition from speech with different systems. This paper attempts emotion recognition from speech which is language independent. The emotional speech samples database is used for feature extraction. For feature extraction MFCC and DWT these two different algorithms are used. For classification of different emotions like angry, happy, scared and neutral state SVM classifier is used. The classification is based on the feature vector formed by fusion of two algorithms. This classified emotion is used for ATM security system.