A hybrid deep learning model combining EfficientNet and Vision Transformers for accurate deepfake image detection. Trained on FF++ and DFDC datasets, the model improves feature extraction, generalization, and precision across manipulated media.

Midterm Assessment for Introduction to Artificial Intelligence Course.

Deepfake Image Detection with Lightweight Models

DeepShield empowers individuals, journalists, and organizations to verify the authenticity of digital media in an age of misinformation.

[NeurIPS'25 Workshop GenProCC] SHIELD: A Benchmark Study on Zero-Shot Detection of AI-Edited Images with Vision Language Models

Detect whether an image of a face is a real person or an AI-generated forgery. Our work builds on the baseline set down by Diffusion Facial Forgery Detection (https://arxiv.org/abs/2401.15859)

AI-powered OSINT platform for image authenticity verification, metadata extraction, and deepfake detection.

Advanced voice cloning and speech synthesis system that can mimic any voice with just 30 seconds of audio - enterprise-grade voice AI with emotion control.

Advanced multi-model deepfake detection framework combining ensemble machine learning with deep neural networks for robust image authentication and media forensics.

Official repository of "Beyond Spatial Frequency: Pixel-wise Temporal Frequency-based Deepfake Video Detection" [ICCV 2025]

Dual-stream deep learning architecture for deepfake detection combining spatial (RGB) and frequency-domain (DCT) analysis with CBAM attention mechanisms.

BWS 3: BWS Client with Web App for Liveness Detection, Video Liveness Detection and Photo ID Verification

Implementation of SSL ViT to Detect DeepFakes, pretrained on CelebA, finetuned on DFDC's train split.

PixelGuard protects images from AI scraping and unauthorized use in AI training, such as facial recognition models or style transfer algorithms. It employs multiple invisible protection techniques that mostly imperceptible to the eye but can interfere with AI processing.

AI-Generated Video Detection via Perceptual Straightening (NeurIPS2025)

A multi-stream deepfake detection model based on spatial, temporal, and facial landmark features, designed to generalize across manipulation types through modular and complementary visual analysis.

FIRM-Net for robust deepfake detection, leveraging PyTorch and EfficientNetV2. Enhances generalization on DFDC

[CVPR'25] CO-SPY: Combining Semantic and Pixel Features to Detect Synthetic Images by AI

A Survey on Deepfake Generation and Detection

Combating Falsification of Speech Videos with Live Optical Signatures (CCS '25)