Combination method of Bone-conduction Speech and Air-conduction Speechfor Speaker Recognition

摘要

Recently, some new sensors, such as bone-conductive micro-phones, throat microphones, and non-audible murmur (NAM) microphones, besides conventional condenser microphones have been developed for collecting speech data. Accordingly, some researchers began to study speaker and speech recognition using speech data collected by these new sensors. We focus on bone-conduction speech data collected by the bone-conductive microphone. This paper proposes a novel speaker identification method which combines "bone-conduction speech" and "air-conduction speech". The proposed method conducts speaker identification by integrating the similarity calculated by air-conduction speech model and similarity calculated by bone-conduction speech model. For evaluating the proposed method, we conduct the speaker identification experiment using part of a large bone-conduction speech corpus constructed by National Research Institute of Police Science, Japan (NRIPS). Experimental results show that the proposed method can reduce a identification error rate of air-conduction speech and bone-conduction speech. Especially, the proposed method achieves that the average error reduction rate from air-conduction speech to the proposed method is 35.8%.