CLIP-Flow: A Universal Discriminator for AI-Generated Images Inspired by Anomaly Detection

TL;DR

This paper introduces CLIP-Flow, a universal anomaly detection method for AI-generated images that leverages pre-trained CLIP features and unsupervised learning, effectively identifying images from unseen generative models without needing AI-generated images for training.

Contribution

The paper presents a novel, generalizable AI-generated image detector using CLIP features and a normalizing flow model trained with proxy images, improving detection of unseen AIIs.

Findings

Effective detection of AIIs from various generators

No need for AI-generated images during training

High generalization to unseen generative models

Abstract

With the rapid advancement of AI generative models, the visual quality of AI-generated images (AIIs) has become increasingly close to natural images, which inevitably raises security concerns. Most AII detectors often employ the conventional image classification pipeline with natural images and AIIs (generated by a generative model), which can result in limited detection performance for AIIs from unseen generative models. To solve this, we proposed a universal AI-generated image detector from the perspective of anomaly detection. Our discriminator does not need to access any AIIs and learn a generalizable representation with unsupervised learning. Specifically, we use the pre-trained CLIP encoder as the feature extractor and design a normalizing flow-like unsupervised model. Instead of AIIs, proxy images, e.g., obtained by applying a spectral modification operation on natural images, are used for training. Our models are trained by minimizing the likelihood of proxy images, optionally combined with maximizing the likelihood of natural images. Extensive experiments demonstrate the effectiveness of our method on AIIs produced by various image generators.