Ultra-lightweight (600KB) Face Anti-Spoofing classifier. Optimized MiniFASNetV2-SE implementation validated on 70k+ samples with ~98% accuracy for edge devices.

Multiple face anti-spoofing models from github

An official implement of "UniVR: A Unified Framework for Pitch-Shifted Voice Restoration in Speaker Identification"

code for the paper "Impact of Channel Variation on One-Class Learning for Spoof Detection"

AI system for detecting GNSS spoofing attacks using physics-aware feature engineering and ML ensemble models.

Syntax machine learning model of Data Analytics Competition of Find IT UGM 2026. This model classifying six face categories while detecting diverse spoofing attacks.

About Computer vision pipeline for real-time spoof detection using object detection, monocular depth estimation, depth variance analysis, and clip.

DNS Lookup Tool with Spoof Detection: A Python script that queries multiple public DNS servers to detect DNS spoofing by comparing resolved IP addresses. Helps enhance DNS security awareness and troubleshooting.

Worked on creating a all in one solution for various sub problems of face detection which includes Blur detection , Professionalism check , Spoof detection , Watermark detection and Obstruction detection.

Computer vision pipeline for real-time spoof detection using object detection, monocular depth estimation, depth variance analysis, and clip.

Built a production-grade electric meter spoof detection model at Sujanix Pvt. Ltd. — multi-branch architecture (MobileNetV2 + FFT + SRM + CBAM), handling 1:18 class imbalance with focal loss. 82.3% accuracy on real-world data. Code on GitHub

Active liveness and deepfake detection pipeline using webcam red/blue light challenges, MediaPipe face landmarks, OpenCV, and temporal-spatial-frequency analysis for face anti-spoofing and presentation attack detection.