Summary

This repository serves as a comprehensive, curated collection of the latest research, datasets, benchmarks, and practical tools for detecting AI-generated images and videos. It targets engineers, researchers, and power users seeking to understand and implement solutions for identifying synthetic media, offering a centralized resource to navigate the rapidly evolving landscape of AIGC detection.

How It Works

The project functions as a living index, meticulously gathering and organizing cutting-edge resources. It categorizes academic contributions into MLLM-based and Classification-based approaches, detailing numerous benchmarks and datasets with specifications like modality, annotation type, and scale. Additionally, it lists practical detection tools and highlights recent industry events, providing a broad overview of the field's advancements and practical applications.

Quick Start & Requirements

This repository is a curated list of resources and does not contain runnable code or direct installation instructions. Users must follow the provided links to access specific datasets, research papers, or external detection tools.

Highlighted Details

• Features an extensive catalog of over 30 benchmarks and datasets (e.g., WildFake, GenVidBench, HydraFake, OpenSDI), detailing their modality, annotation type, scale, and generative sources.
• Compiles a significant number of recent research papers (primarily 2025-2026) focused on both MLLM-based and traditional classification-based detection methods, with many including links to code repositories.
• Includes a section on "Hot Events" summarizing key developments and discussions in the AIGC detection space, alongside