Effective speaker verification via dynamic mismatch compensation

Abstract

This paper presents a new approach to Condition-adjusted T-Norm (CT-Norm) for speaker verification under significant mismatched noise conditions. The study is motivated by the fact that, whilst the standard CT-Norm method offers enhanced accuracy under mismatched data conditions, its effectiveness reduces with the increased severity of such conditions. The proposed approach attempts to address this challenge by providing a more effective reduction of data mismatch through the incorporation of multi-SNR UBMs (universal background models). The effectiveness of the proposed approach is demonstrated through experiments based on examples of real-world noise. It is shown that the superiority of the approach over CT-Norm is particularly significant for such excessive levels of test data degradation considered in the study as 5 dB and below. The paper provides a description of the characteristics of the proposed approach and details the experimental analysis of its effectiveness under different noise conditions

Open-set speaker identification with diverse-duration speech data

“…These approaches are based on firm pattern matching principles, and incorporate capabilities for dealing with the effects of noise and other causes of variation in speech characteristics 2,3 . However, to date, there has been limited attention to the challenging problems posed by operating under uncontrolled conditions in certain applications.…”

Voice biometrics using linear Gaussian model

“…A voice signal contains information about a person's physiological characteristics such as the vocal chords, glottis and vocal tract dimensions. Voice biometrics (speaker recognition) is a biometric method that encompasses identification and verification through voice signal processing [1]. In particular, voice biometrics often works together with other biometrics methods, such as fingerprint recognition [2], face recognition [3], ear recognition [4] and so on.…”