UTRAD: Anomaly detection and localization with U-Transformer

Anomaly detection is an active research field in industrial defect detection and medical disease detection. However, previous anomaly detection works suffer from unstable training, or non-universal criteria of evaluating feature distribution. In this paper, we introduce UTRAD, a U-TRansformer based Anomaly Detection framework. Deep pre-trained features are regarded as dispersed word tokens, and represented with transformer-based autoencoders. With reconstruction on more informative feature distribution instead of raw images, we achieve a more stable training process and a more precise anomaly detection and localization result. In addition, our proposed UTRAD has a multi-scale pyramidal hierarchy with skip connections that help detect both multi-scale structural and non-structural anomalies. As attention layers are decomposed to multi-level patches, UTRAD significantly reduces the computational cost and memory usage compared with the vanilla transformer. Experiments on industrial dataset MVtec AD and medical datasets Retinal-OCT, Brain-MRI, Head-CT have been conducted. Our proposed UTRAD out-performs all other state-of-the-art methods in the above datasets. Code released at https://github.com/gordon-chenmo/UTRAD. (c) 2021 Elsevier Ltd. All rights reserved.

Automated summary

This paper introduces UTRAD, a U-Transformer based Anomaly Detection framework that achieves a more stable training process and precise anomaly detection and localization results by reconstructing more informative feature distribution instead of raw images. UTRAD also reduces computational cost and memory usage with a multi-scale pyramidal hierarchy and skip connections.