One-Step Diffusion with Inverse Residual Fields for Unsupervised Industrial Anomaly Detection

TL;DR

This paper introduces OSD-IRF, a one-step diffusion method for unsupervised industrial anomaly detection that improves speed and effectiveness by analyzing inverse residual fields.

Contribution

The paper proposes a novel one-step diffusion approach using inverse residual fields for faster and more accurate unsupervised industrial anomaly detection.

Findings

Achieves state-of-the-art or competitive performance on three benchmarks.

Provides roughly 2X inference speedup without distillation.

Identifies anomalies in the inverse residual field space.

Abstract

Diffusion models have achieved outstanding performance in unsupervised industrial anomaly detection (uIAD) by learning a manifold of normal data under the common assumption that off-manifold anomalies are harder to generate, resulting in larger reconstruction errors in data space or lower probability densities in the tractable latent space. However, their iterative denoising and noising nature leads to slow inference. In this paper, we propose OSD-IRF, a novel one-step diffusion with inverse residual fields, to address this limitation for uIAD task. We first train a deep diffusion probabilistic model (DDPM) on normal data without any conditioning. Then, for a test sample, we predict its inverse residual fields (IRF) based on the noise estimated by the well-trained parametric noise function of the DDPM. Finally, uIAD is performed by evaluating the probability density of the IRF under a Gaussian distribution and comparing it with a threshold. Our key observation is that anomalies become distinguishable in this IRF space, a finding that has seldom been reported in prior works. Moreover, OSD-IRF requires only single step diffusion for uIAD, thanks to the property that IRF holds for any neighboring time step in the denoising process. Extensive experiments on three widely used uIAD benchmarks show that our model achieves SOTA or competitive performance across six metrics, along with roughly a 2X inference speedup without distillation.