media-forensics

Image Forgery Detection and Localization (and related) Papers List

Agent skill for deepfake detection & media safety — detect AI-generated audio, images, and video with Resemble AI

AutoSplice: A Text-prompt Manipulated Image Dataset for Media Forensics, WMF@CVPR2023

This GitHub provides different DeepFakes Detectors using facial regions and considering three different state-of-the-art fake detection systems.

Test what AI watermark and provenance evidence exists, whether it verifies, and what survives publishing.

This repository contains the official implementation (PyTorch) of "Multimodal Forgery Detection Using Ensemble Learning" proposed in APSIPA Paper 2022.

Deep learning system achieving 95.36% accuracy in media forgery detection using hybrid ResNet50+ViT architecture. Optimized for 20% training data efficiency with PyTorch, OpenCV, and Flask-based inference pipeline. Live demo on Hugging Face Spaces.

Three forensic worlds. One conversation. Zero paperwork

Production-ready Multimodal Lip Sync Detection & Deepfake Detection System. Detects audio-video synchronization mismatches using deep learning (PyTorch) with a scalable FastAPI-based inference pipeline. Optimized for real-time processing,low false positives, and robust performance on noisy speech segments. Built for video forensics,synthetic media

Checks whether a news link, image or short video is what it claims to be. Local forensics plus an LLM fact-check, with the claims and sources shown.

🎭🔍 Detecting deepfake videos and images using Deep Learning with a Tkinter desktop interface

AI-powered deepfake detection system using Deep Learning and Computer Vision to identify manipulated facial images and videos with high accuracy.

M.Eng. Thesis (NWPU): Detecting Deepfake Video by a Multimodal Audio-Visual Framework with Temporal Inconsistencies — EfficientNet+FFT, MFCC, and Transformer attention fusion on WildDeepfake.

Enterprise Generative Media Provenance & Forensics Suite. Detects tampering and tracks AI asset origin using robust, imperceptible frequency-domain watermarking and signed C2PA manifests.

InSwapper Detector is a production-focused deepfake detection system designed to identify faces manipulated with INSwapper. It combines face detection, RGB image analysis, frequency artifact extraction, and a multi-task ConvNeXt-Tiny model for image, batch, and video-based detection.

AI-powered multimodal deepfake detection system leveraging audio-visual fusion, transformer architectures, physiological analysis, and forensic intelligence for robust media authenticity verification.

Dual-stream spectral-spatial neural network for detecting AI-generated images (CIFAKE). 95.71% accuracy, 13.2x smaller than ResNet-18, with a full two-run replication that refutes the project's own central hypothesis.

CNN-based video forgery (deepfake) detection on public data, with embedding analysis and threshold tuning from ROC/PR curves.

Real-time face-swap artifact detector — webcam/mobile frames + video jobs (OpenCV, PyTorch, Flask)

An Agentic AI system for Deepfake Detection & Media Authenticity Verification. Features autonomous model routing, an ensemble of 6 specialized neural networks, and advanced bias correction mechanisms.