A Unified Model for Multi-class Anomaly Detection

TL;DR

UniAD introduces a unified multi-class anomaly detection model that overcomes common shortcut issues with innovative layer design, neighbor attention, and feature jittering, achieving state-of-the-art results on multiple datasets.

Contribution

The paper presents a novel unified framework for multi-class anomaly detection that incorporates query decoders, neighbor masked attention, and feature jittering to improve detection accuracy.

Findings

Surpasses state-of-the-art on MVTec-AD and CIFAR-10 datasets.

Achieves 96.5% anomaly detection accuracy on 15 categories in MVTec-AD.

Significantly improves anomaly localization performance.

Abstract

Despite the rapid advance of unsupervised anomaly detection, existing methods require to train separate models for different objects. In this work, we present UniAD that accomplishes anomaly detection for multiple classes with a unified framework. Under such a challenging setting, popular reconstruction networks may fall into an "identical shortcut", where both normal and anomalous samples can be well recovered, and hence fail to spot outliers. To tackle this obstacle, we make three improvements. First, we revisit the formulations of fully-connected layer, convolutional layer, as well as attention layer, and confirm the important role of query embedding (i.e., within attention layer) in preventing the network from learning the shortcut. We therefore come up with a layer-wise query decoder to help model the multi-class distribution. Second, we employ a neighbor masked attention module to further avoid the information leak from the input feature to the reconstructed output feature. Third, we propose a feature jittering strategy that urges the model to recover the correct message even with noisy inputs. We evaluate our algorithm on MVTec-AD and CIFAR-10 datasets, where we surpass the state-of-the-art alternatives by a sufficiently large margin. For example, when learning a unified model for 15 categories in MVTec-AD, we surpass the second competitor on the tasks of both anomaly detection (from 88.1% to 96.5%) and anomaly localization (from 89.5% to 96.8%). Code is available at https://github.com/zhiyuanyou/UniAD.