One-Class Learning with Adaptive Centroid Shift for Audio Deepfake Detection

arXiv:2406.16716v1 Announce Type: cross
Abstract: As speech synthesis systems continue to make remarkable advances in recent years, the importance of robust deepfake detection systems that perform well in unseen systems has grown. In this paper, we propose a novel adaptive centroid shift (ACS) method that updates the centroid representation by continually shifting as the weighted average of bonafide representations. Our approach uses only bonafide samples to define their centroid, which can yield a specialized centroid for one-class learning. Integrating our …

arxiv audio audio deepfake class continue cs.cr cs.sd deepfake deepfake detection detection eess.as novel remarkable representation speech synthesis systems updates