Noise-Robust Speaker Recognition Combining Missing Data Techniques and Universal Background Modeling

• Published in: IEEE transactions on audio, speech, and language processing Vol. 20; no. 1; pp. 108 - 121
• Main Authors: May, T., van de Par, S., Kohlrausch, A.
• Format: Journal Article
• Language: English
• Published: Piscataway, NJ IEEE 01.01.2012; Institute of Electrical and Electronics Engineers
• Subjects: Adaptation model; Applied sciences; Automatic speaker recognition (ASR); Data models; Estimation; Exact sciences and technology; Information, signal and communications theory; mask estimation; Materials; mel frequency cepstral coefficient (MFCC); Miscellaneous; missing data; noise robustness; Signal processing; Speaker recognition; Speech; Speech processing; Speech recognition; Telecommunications and information theory; universal background model (UBM); Audio signal processing; Background; mel frequency cepstral coefficient (MFCC); universal background model (UBM); Non stationary condition; Speaker adaptation; Speaker recognition; Acoustic signal processing; Background noise; Modeling; Missing data; Automatic speaker recognition (ASR); Cepstral analysis; Noise immunity; Robustness; Automatic recognition; mask estimation; Speech processing; noise robustness

Although the field of automatic speaker recognition (ASR) has been the subject of extensive research over the past decades, the lack of robustness against background noise has remained a major challenge. This paper describes a noise-robust speaker recognition system that combines missing data (MD) r...