Discriminative feature based on FWMW for playback speech detection. Issue 15 (1st July 2019)

• Record Type: Journal Article
• Authors: Yang, Jichen; Liu, Leian; He, Qianhua
• Is Part Of: Electronics letters. Volume 55:Issue 15(2019)
• Page Start: 861
• Page End: 864
• Publication Date: 2019-07-01
• Subjects: feature extraction -- neural nets -- Gaussian processes; genuine speech -- FWW -- frame magnitude spectrum frequency bin -- discriminative ability -- commonly used features -- playback speech detection -- discriminative feature extraction method -- frame‐wise weight; Electronics -- Periodicals; 621.381
• DOI: 10.1049/el.2019.1025
• Languages: English
• ISSNs: 0013-5194

Abstract

Abstract : A discriminative feature extraction method for playback speech detection is proposed, it relies on the finding that frame‐wise magnitude‐spectrum weight (FWMW) can enlarge the difference between genuine speech and playback speech. The proposed FWMW is obtained by supplying the sum of every frame magnitude spectrum frequency bin as the weight on the frame magnitude spectrum. Then, a new feature based on FWMW is proposed, namely constant‐Q weight segmentation coefficients (CQWSCs). The experimental result on ASVspoof 2017 version 2.0 evaluation set using CQWSC indicates that: (i) FWMW can make the proposed feature has more discriminative ability and the equal error rate can decline 15.53 and 19.18% under Gaussian mixture model and deep neural network, respectively, and (ii) CQWSC performs better than some commonly used features.