ARCHITECTURAL DISCLOSURES
Filed Indian Patents
Public Technical Overviews of Streaming Anomaly Detection, Audio Deepfake Forensics, and Cognitive Security.
Pending / Under Review (Indian Patent Office)
Information Security, Deepfake Detection, Adaptive UI
Visuals reflect high-level conceptual pipelines. Enabling equations, exact thresholds, and implementation internals are omitted.
Cognitive-State-Aware Adaptive Information Security
A non-biometric, content-agnostic middleware architecture that processes user interaction dynamics (typing, navigation, rhythm) to infer cognitive load and vulnerability state, dynamically escalating security friction.
Multi-Domain Feature Fusion for Robust Audio Deepfake Detection
An architecture designed for robustness to acoustic noise and unseen manipulation conditions using multi-domain representation learning across Spectral, Cepstral, and Raw Temporal domains.
Multi-Feature Fusion for Lip-Sync Authenticity Detection
A dual-path forensic architecture analyzing visual-only data (no audio required) by combining a learned visual attention CNN with handcrafted spatial and spectral forensic analysis to detect desynchronization artifacts.
