Code Compilation for AIGC-based Image Detection Algorithm Papers

The explosive boom of Artificial Intelligence Generated Content (AIGC) has led to a grave trust crisis, making AI-generated image detection a central issue in multimedia forensics. To tackle the problems of fragmented research perspectives and disjointed theoretical frameworks in existing detection technologies, this paper proposes a unified analytical framework based on "learning paradigms" for a systematic review of relevant techniques. It constructs a dual-paradigm system of "falsification" and "verification", elaborates on their technical principles, reveals the evolution trend from passive trace detection to active distribution modeling, and analyzes their respective limitations and challenges. Additionally, a multi-dimensional evaluation system is established to compare representative algorithms, with related open-source resources and datasets shared on the ScienceDB Community for future research.