InvAD: Inversion-based Reconstruction-Free Anomaly Detection with Diffusion Models

ArXi:2504.05662v5 Announce Type: replace Despite the remarkable success, recent reconstruction-based anomaly detection (AD) methods via diffusion modeling still involve fine-grained noise-strength tuning and computationally expensive multi-step denoising, leading to a fundamental tension between fidelity and efficiency. In this paper, we propose InvAD, a novel inversion-based anomaly detection approach ("detection via noising in latent space") that circumvents explicit reconstruction.