Robust Speech Hash Function

ETRI Journal

关键词: non-negative matrix factorization (NMF);    linear prediction coefficients (LPCs);    Speech hash function;   

Others  :  1185979
DOI  :  10.4218/etrij.10.0209.0309

【 摘 要 】

In this letter, we present a new speech hash function based on the non-negative matrix factorization (NMF) of linear prediction coefficients (LPCs). First, linear prediction analysis is applied to the speech to obtain its LPCs, which represent the frequency shaping attributes of the vocal tract. Then, the NMF is performed on the LPCs to capture the speech’s local feature, which is then used for hash vector generation. Experimental results demonstrate the effectiveness of the proposed hash function in terms of discrimination and robustness against various types of content preserving signal processing manipulations.

   

【 参考文献 】

• [1]P. Cano et al., "A Review of Audio Fingerprinting," J. VLSI Signal Process., vol. 41, no. 3, 2005, pp. 271-284.
• [2]A. Ramalingam and S. Krishnan, "Gaussian Mixture Modeling of Shorttime Fourier Transform Features for Audio Fingerprinting," IEEE Trans. Inf. Forensics Security, vol. 1, no. 4, 2006, pp. 457-463.
• [3]M. Park, H. Kim, and S.H. Yang, "Frequency-Temporal Filtering for a Robust Audio Fingerprinting Scheme in Real-Noise Environments," ETRI J., vol. 28, no. 4, 2006, pp. 509-512.
• [4]Y. Jiao et al., "Key-Dependent Compressed Domain Audio Hashing," Proc. ISDA, 2008.
• [5]Y. Jiao, Q. Li, and X. Niu, "Compressed DomainPerceptual Hashing for MELP Coded Speech," Proc. IIHMSP, 2008, pp. 410-413.
• [6]D.D. Lee and H.S. Seung, "Learning the Parts of Objects by Non-negative Matrix Factorization,"Nature, vol. 401, no. 6755, 1999, pp. 788-791.