Synthetic Image Verification in the Era of Generative AI: What Works and What Isn't There Yet

TL;DR

This paper reviews current methods for detecting and attributing synthetic images generated by AI, discussing their effectiveness, limitations, and future research directions in the rapidly evolving field.

Contribution

It provides a comprehensive overview of existing approaches, analyzing their strengths and weaknesses, and identifies promising future research directions in synthetic image verification.

Findings

Current detection methods have varying effectiveness.

Many approaches face challenges with advanced generative models.

Future research should focus on robust, scalable verification techniques.

Abstract

In this work we present an overview of approaches for the detection and attribution of synthetic images and highlight their strengths and weaknesses. We also point out and discuss hot topics in this field and outline promising directions for future research.