AegisVox: Factorized Privacy and Clinical Preservation for Long-Form Speech
Clinical speech reveals identity through both spoken identifiers and voice, yet its pauses, delivery, and speaker roles can also carry useful clinical evidence. Protecting that speech therefore requires more than preserving its words. We present AegisVox, a factorized framework that assigns consistent surrogate voices...
Praphul Singh, Akshat Agarwal, Ratul Ghosh et al.
· 0 citations