Forensic Lens: Deepfake Detection Through Micro-Level Facial Blood-Flow Signals
Hassan Minhal Raza
Abstract & Problem Statement
Synthetic facial manipulation techniques and generative adversarial networks have evolved rapidly, creating significant threats to digital identity and forensic evidence authenticity. This research introduces Forensic Lens, an approach that monitors subtle micro-level chromatic fluctuations in facial epidermal tissue induced by cardiac cardiovascular cycles (photoplethysmography). Because generative algorithms fail to accurately reproduce spatio-temporal biological blood flow coherently across localized facial sub-regions, our framework exposes synthetic manipulations with strong discriminative power.