Sub-Image Anomaly Detection with Deep Pyramid Correspondences

TL;DR

This paper introduces SPADE, a novel anomaly segmentation method that aligns images with similar normal images using multi-resolution features, achieving state-of-the-art results with minimal training.

Contribution

SPADE is a new approach that provides both anomaly detection and localization by aligning images with normal counterparts using deep pyramid correspondences.

Findings

Achieves state-of-the-art anomaly detection and localization performance.

Requires virtually no training time.

Effective in unsupervised anomaly detection tasks.

Abstract

Nearest neighbor (kNN) methods utilizing deep pre-trained features exhibit very strong anomaly detection performance when applied to entire images. A limitation of kNN methods is the lack of segmentation map describing where the anomaly lies inside the image. In this work we present a novel anomaly segmentation approach based on alignment between an anomalous image and a constant number of the similar normal images. Our method, Semantic Pyramid Anomaly Detection (SPADE) uses correspondences based on a multi-resolution feature pyramid. SPADE is shown to achieve state-of-the-art performance on unsupervised anomaly detection and localization while requiring virtually no training time.