Noise Clustering Approach to Speaker Verification

In a speaker veri ̄cation system, a claimed speaker's score is computed to accept or reject the speaker claim. Most of the current methods compute the score as the ratio of the claimed speaker's and the impostors' likelihood functions. Based on analysing false acceptance error obtained by using these methods, we propose a noise clustering approach to ̄nd better scores which can reduce that error. The noise clustering method was used to deal with noisy data for fuzzy clustering methods. In speaker veri ̄cation, impostor's utterances are considered as noisy data and thus should have arbitrarily small fuzzy membership functions in the claimed speaker's fuzzy subset. Experiments performed on the ANDOSL and YOHO speech corpora show better results for the proposed method.