OmiAD: One-Step Adaptive Masked Diffusion Model for Multi-class Anomaly Detection via Adversarial Distillation

Diffusion models have demonstrated outstanding performance in industrial anomaly detection. However, their iterative denoising nature results in slow inference speed, limiting their practicality for real-time industrial deployment. To address this challenge, we propose OmiAD, a one-step masked diffusion model for multi-class anomaly detection, derived from a well-designed multi-step Adaptive Masked Diffusion Model (AMDM) and compressed using Adversarial Score Distillation (ASD). OmiAD first introduces AMDM, equipped with an adaptive masking strategy that dynamically adjusts masking patterns based on noise levels and encourages the model to reconstruct anomalies as normal counterparts by leveraging broader context, to reduce the pixel-level shortcut reliance. Then, ASD is developed to compress the multi-step diffusion process into a single-step generator by score distillation and incorporating a shared-weight discriminator effectively reusing parameters while significantly improving both inference efficiency and detection performance. The effectiveness of OmiAD is validated on four diverse datasets, achieving state-of-the-art performance across seven metrics while delivering a remarkable inference speedup.

Detecting unusual patterns or defects in images—like scratches on a smartphone screen or flaws in manufactured goods—is important in many real-world applications. Our work introduces a new method called OmiAD that helps computers find such defects faster and more accurately. OmiAD teaches a small, fast model to behave like a larger, more accurate one, using a process inspired by how humans learn from teachers. This makes it possible to spot problems in real time, which is especially useful in industrial settings like factories. Our method works well across different types of defects and offers a practical solution for automated quality inspection.