SPEAKER RECOGNITION VIA SPARSE REPRESENTATIONS USING ORTHOGONAL MATCHING PURSUIT

The objective of this paper is to demonstrate the effectiveness of sparse representation techniques for speaker recognition. In this ap-proach, each feature vector from unknown utterance is expressed as linear weighted sum of a dictionary of feature vectors belonging to many speakers. The weights associated with feature vectors in the dictionary are evaluated using orthogonal matching pursuit al-gorithm, which is a greedy approximation to l0 optimization. The weights thus obtained exhibit high level of sparsity, and only a few of them will have nonzero values. The feature vectors which be-long to the correct speaker carry significant weights. The proposed method gives an equal error rate (EER) of 10.84 % on NIST-2003 database, whereas the existing GMM-UBM system gives an EER of 9.67%. By combining evidence from both the systems an EER of 8.15 % is achieved, indicating that both the systems carry compli-mentary information. Index Terms — Sparse representation, orthogonal matching pur-suit, l0 optimization and Gaussian mixture modeling. 1