Research And Design Of Voiceprint Recognition System For Specific Human Group

Voiceprint recognition technology is the applied technology that distinguishes or automatically recognizes speaker's identity on the basis of physiological and behavioral feature included in speech signals. The application of voiceprint recognition technology provide a more convenient and reliable method for identity recognition in modern information society, so it's a very important applied significance to research and develop the voiceprint recognition system.The voiceprint recognition system is generally made up of three main functional modules including speech pretreatment, feature parameter extraction and pattern matching judgment. In this paper, based on the analysis of the basic principles and key technologies of voiceprint recognition technology, a design of voiceprint recognition for specific human group is presented, then the three main functional modules are introduced in detail, and with the experiment result analysis, it proves the practicality and feasibility of the system design.The innovation in this paper is that: Linear Prediction Cepstrum Coefficient(LPCC)N Mel Frequency Cepstrum Coefficient(MFCC) with its difference Cepstrum Coefficient have been combined as feature parameters together. It can improve the accuracy of speech feature parameter extraction, and it combines with DTW and VQ recognition methods to show the speaker's individual feature including the speaker's track feature, hearing feature and speech dynamic movement. So it can efficiently improve the recognition rate in the voiceprint recognition system.