APPLYING MFCC-BASED AUTOMATIC SPEAKER RECOGNITION TO GSM AND FORENSIC DATA
暂无分享,去创建一个
Speaker Profiler computer program for automatic speaker recognition has been developed in a research project funded by the Finnish Technology Agency. A vector quantization (VQ) matching approach is used, where dissimilarity of an unknown speech sample is computed for codebooks created using the K-means algorithm. This study tests the recognition reliability with two databases constructed from Finnish band-limited GSM speech and authentic crime case speech. Material for the first test is recorded with a GSM phone and a laptop computer. Spontaneous speech vs. reading was tested. The program should pick the right person from the database based on independent non-verbatim speech samples. There were 47.5 % out of 107 samples ranked first correctly. Some very poor quality speech files were used in training and the mother tongue for some speakers was not Finnish. If these samples were not considered, the result was better. The second part of this study consists of real crime investigation cases. The speaker database was constructed from known speech samples (suspect). Unknown sample(s) recorded at the crime scene were matched against the database. From the matched 61 samples, 68.9 % were ranked first correctly. Accuracy is sufficient for creating shortlists in forensics.
[1] Tomi Kinnunen,et al. Spectral Features for Automatic Text-Independent Speaker Recognition , 2003 .
[2] Tomi Kinnunen,et al. Real-time speaker identification and verification , 2006, IEEE Transactions on Audio, Speech, and Language Processing.
[3] P. Fränti,et al. Fusion of Spectral Feature Sets for Accurate Speaker Identification , 2004 .
[4] F Botti,et al. On compensation of mismatched recording conditions in the Bayesian approach for forensic automatic speaker recognition. , 2004, Forensic science international.