A Novel Approach To Enhance The Efficiency Of Real-Time Speaker Identification System

Abstract

This paper gives overall information about a real-time speaker identification system (SI). We create a real-time database taking all different real-time conditions. We use baseline cepstral features, i.e., Mel Frequency Cepstral Coefficients (MFCC) alone with proposed approach, i.e., Speech signal-based Frequency Cepstral Coefficients (SFCC). Here the proposed system provides better performance than that of MFCC for real time speaker identification system. The overall identification accuracy of proposed approach is 27% relative improvement over baseline method.