Search papers, labs, and topics across Lattice.
This paper introduces a training-free, search-based method for detecting AI-generated images by analyzing the sensitivity of image embeddings in intermediate layers of a pre-trained model to small perturbations. The method compares the similarity between original and perturbed image embeddings, using the degree of similarity to classify images as real or AI-generated. Experiments on GenImage and Forensics Small benchmarks demonstrate that this approach outperforms both training-free and training-based state-of-the-art methods, achieving a 39.61% AUROC improvement over the best training-free method on Forensics Small.
Forget retraining: probing intermediate layer embedding sensitivity offers a surprisingly effective and efficient way to spot AI-generated images.
The rapid advancement in generative AI models has enabled the creation of photorealistic images. At the same time, there are growing concerns about the potential misuse and dangers of generated content, as well as a pressing need for effective AI-generated image detectors. However, current training-based detection techniques are typically computationally costly and can hardly be generalized to unseen data domains, while training-free methods fall short in detection performance. To bridge this gap, we propose a search-based method employing data embedding sensitivity in intermediate layers to detect AI-generated images. Given a set of real and AI-generated images, our method examines the similarity between original image embeddings and perturbed image embeddings, and detects AI-generated images based on the similarity. We examine the proposed method on two comprehensive benchmarks: GenImage and Forensics Small. Our method exhibits improved performance across different datasets compared to both training-free and training-based state-of-the-art methods. On average, our method achieves the largest performance gain on the Forensics Small benchmark by 39.61% compared to the best training-free method and 5.14% compared to the best training-based method in AUROC score.