Testing Voice Mimicry with the YOHO Speaker Verification Corpus
نویسندگان
چکیده
The aim of this paper is to determine how vulnerable a speaker verification system is to conscious effort by impostors to mimic a client of the system. The paper explores systematically how much closer an impostor can get to another speakers voice by repeated attempts. Experiments on 138 speakers in the YOHO database and six people who played a role as imitators showed a fact that professional linguists could successfully attack the system. Nonprofessional people could have a good chance if they know their closest speaker in the database.
منابع مشابه
I-vectors meet imitators: on vulnerability of speaker verification systems against voice mimicry
Voice imitation is mimicry of another speaker’s voice characteristics and speech behavior. Professional voice mimicry can create entertaining, yet realistic sounding target speaker renditions. As mimicry tends to exaggerate prosodic, idiosyncratic and lexical behavior, it is unclear how modern spectral-feature automatic speaker verification systems respond to mimicry “attacks”. We study the vul...
متن کاملSpeaker characterization using principal component analysis and wavelet transform for speaker verification
In this paper, we investigate the use of the Wavelet Transform for text-dependent and text-independent Speaker Verification tasks. We have introduced a Principal Component Analysis based wavelet transform to perform frequencies segmentation with levels decomposition. A speaker dependent library tree has been built, corresponding to the best structure for a given speaker. The constructed tree is...
متن کاملComparison of Human Listeners and Speaker Verification Systems Using Voice Mimicry Data
In this work, we compare the performance of human listeners and two well known speaker verification systems in presence of voice mimicry. Our focus is to gain insights on how well human listeners recognize speakers when mimicry data is included and compare it to the overall performance of state-ofthe-art speaker verification systems, a traditional Gaussian mixture model-universal background mod...
متن کاملAn Attack on a Text-independent Speaker Authentication System
We mount an effective attack on a third-party open-source text-independent speaker verification system. Specifically, we show how an attacker can simply use a signal generated at fixed frequency to pass speaker verification and gain access to other user accounts. We demonstrate this attack on the GMM-UBM based ALIZE speaker verification system using the YOHO voice database. We show through expe...
متن کاملTitle Speaker Recognition Using Adaptively Boosted Decision Tree Classifier( Accepted Version )
In this paper, a novel approach for speaker recognition is proposed. The approach makes use of adaptive boosting (AdaBoost) and C4.5 decision trees for closed set, text-dependent speaker recognition. A subset of 20 speakers, 10 male and 10 female, drawn from the YOHO speaker verification corpus is used to assess the performance of the system. Results reveal that an accuracy of 99.5% of speaker ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2005