Sub-Image Anomaly Detection with Deep Pyramid Correspondences
This work addresses the need for precise anomaly localization in images, which is crucial for applications like industrial inspection or medical imaging, representing a novel approach to an existing bottleneck.
The paper tackled the problem of anomaly segmentation in images by introducing SPADE, a method that uses multi-resolution feature pyramids to align anomalous images with normal ones, achieving state-of-the-art performance in unsupervised anomaly detection and localization with minimal training time.
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.