Subspace-Guided Feature Reconstruction for Unsupervised Anomaly Localization

TL;DR

This paper introduces a subspace-guided feature reconstruction method for unsupervised anomaly localization, improving robustness and efficiency by modeling features with learned low-dimensional subspaces.

Contribution

It proposes a novel framework that constructs low-dimensional subspaces from nominal samples and reconstructs target features via a self-expressive model, enhancing anomaly detection.

Findings

Achieves state-of-the-art performance on three benchmark datasets.

Reduces memory overhead through a sparsity-based sampling method.

Effectively models out-of-bank features for better anomaly localization.

Abstract

Unsupervised anomaly localization aims to identify anomalous regions that deviate from normal sample patterns. Most recent methods perform feature matching or reconstruction for the target sample with pre-trained deep neural networks. However, they still struggle to address challenging anomalies because the deep embeddings stored in the memory bank can be less powerful and informative. Specifically, prior methods often overly rely on the finite resources stored in the memory bank, which leads to low robustness to unseen targets. In this paper, we propose a novel subspace-guided feature reconstruction framework to pursue adaptive feature approximation for anomaly localization. It first learns to construct low-dimensional subspaces from the given nominal samples, and then learns to reconstruct the given deep target embedding by linearly combining the subspace basis vectors using the self-expressive model. Our core is that, despite the limited resources in the memory bank, the out-of-bank features can be alternatively ``mimicked'' to adaptively model the target. Moreover, we propose a sampling method that leverages the sparsity of subspaces and allows the feature reconstruction to depend only on a small resource subset, contributing to less memory overhead. Extensive experiments on three benchmark datasets demonstrate that our approach generally achieves state-of-the-art anomaly localization performance.