Detecting Diffusion-generated Images via Dynamic Assembly Forests

TL;DR

This paper introduces a novel, resource-efficient machine learning model called Dynamic Assembly Forest (DAF) for detecting images generated by diffusion models, offering a practical alternative to deep neural networks.

Contribution

The paper proposes DAF, a deep forest-based model that effectively detects diffusion-generated images with fewer parameters and lower computational costs than traditional DNNs.

Findings

DAF achieves competitive detection performance.

DAF has significantly fewer parameters and lower computational cost.

Code and models are publicly available at https://github.com/OUC-VAS/DAF.

Abstract

Diffusion models are known for generating high-quality images, causing serious security concerns. To combat this, most efforts rely on deep neural networks (e.g., CNNs and Transformers), while largely overlooking the potential of traditional machine learning models. In this paper, we freshly investigate such alternatives and proposes a novel Dynamic Assembly Forest model (DAF) to detect diffusion-generated images. Built upon the deep forest paradigm, DAF addresses the inherent limitations in feature learning and scalable training, making it an effective diffusion-generated image detector. Compared to existing DNN-based methods, DAF has significantly fewer parameters, much lower computational cost, and can be deployed without GPUs, while achieving competitive performance under standard evaluation protocols. These results highlight the strong potential of the proposed method as a practical substitute for heavyweight DNN models in resource-constrained scenarios. Our code and models are available at https://github.com/OUC-VAS/DAF.