Reproduce “SENTRY: A Streaming Edge Network Trust Engine for Real-Time Audio Deepfake Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Chirantan Banerjee, Soumita Chatterjee, Bhargab B. Bhattacharya, Hafizur Rahaman
SENTRY is an edge-oriented voice-phishing detector that turns streaming audio into integer-based acoustic features and combines supervised spoof classification with an authentic-speech out-of-distribution score. A smoothed fused score selects LOG, CHALLENGE, or DROP. On closed-set WaveFake and ASVspoof 2019 LA tests, it reports 98.75% and 94.75% accuracy; unseen generators and end-to-end call verdict time remain unmeasured.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.