The current mainstream and state-of-the-art anomaly detection (AD) methods are substantially established on pretrained feature networks yielded by ImageNet pre- training. However, regardless of supervised or self-supervised pretraining, the pretraining process on ImageNet does not match the goal of anomaly detection (i.e., pretraining in natural images doesn’t aim to distinguish between normal and abnormal). Moreover, natural images and industrial image data in AD scenarios typically have the distribution shift. The two issues can cause ImageNet-pretrained features to be suboptimal for AD tasks. To further promote the development of the AD field, pretrained representations specially for AD tasks are eager and very valuable. To this end, we propose a novel AD representation learning framework specially designed for learning robust and discriminative pretrained representa- tions for industrial anomaly detection. Specifically, closely surrounding the goal of anomaly detection (i.e., focus on discrepancies between normals and anoma- lies), we propose angle- and norm-oriented contrastive losses to maximize the angle size and norm difference between normal and abnormal features simulta- neously. To avoid the distribution shift from natural images to AD images, our pretraining is performed on a large-scale AD dataset, RealIAD. To further alle- viate the potential shift between pretraining data and downstream AD datasets, we learn the pretrained AD representations based on the class-generalizable repre- sentation, residual features. For evaluation, based on five embedding-based AD methods, we simply replace their original features with our pretrained represen- tations. Extensive experiments on five AD datasets and five backbones consis- tently show the superiority of our pretrained features. The code is available at https://github.com/xcyao00/ADPretrain.

This paper explores the problem of class-generalizable anomaly detection, where the objective is to train one unified AD model that can generalize to detect anomalies in diverse classes from different domains without any retraining or fine-tuning on the target data. Because normal feature representations vary significantly across classes, this will cause the widely studied one-for-one AD models to be poorly classgeneralizable (i.e., performance drops dramatically when used for new classes). In this work, we propose a simple but effective framework (called ResAD) that can be directly applied to detect anomalies in new classes. Our main insight is to learn the residual feature distribution rather than the initial feature distribution. In this way, we can significantly reduce feature variations. Even in new classes, the distribution of normal residual features would not remarkably shift from the learned distribution. Therefore, the learned model can be directly adapted to new classes. ResAD consists of three components: (1) a Feature Converter that converts initial features into residual features; (2) a simple and shallow Feature Constraintor that constrains normal residual features into a spatial hypersphere for further reducing feature variations and maintaining consistency in feature scales among different classes; (3) a Feature Distribution Estimator that estimates the normal residual feature distribution, anomalies can be recognized as out-of-distribution. Despite the simplicity, ResAD can achieve remarkable anomaly detection results when directly used in new classes. The code is available at https://github.com/xcyao00/ResAD.

Unified (multi-class and cross-class) anomaly detection (AD) is a growing area of interest in real-world applications. However, the popular reconstruction-based AD approach usually faces two significant challenges: the "identical shortcut" issue (copying the input as output) and the lack of class adaptability (the AD model cannot be directly applied to new classes). To address these challenges, we propose a novel unified AD method, named IRAD (Input-Reference Joint Driven). Our core insight is to effectively incorporate both input and references into the reconstruction process. Our IRAD consists of three components: 1) A Suspicious Anomaly Substituting module that replaces the potential abnormal regions of input with anomaly-free reference patches to prevent abnormal information leakage, effectively addressing the "identical shortcut". 2) An Input-Reference Fusing module that merges reference embeddings with input, which urges the subsequent Decoder to effectively utilize the normal reference patterns to reconstruct anomaly-free samples, making our model more class-adaptive. 3) A Rich Feature Preserving Decoder that efficiently preserves low-level details, mitigating low-level information degradation during reverse construction from high to low level. In multi-class AD, IRAD achieves better results on Mvtec-AD, BTAD, and VisA. In cross-class AD, IRAD also outperforms the baesline methods on Mvtec-AD and VisA.

One of the most challenges for anomaly detection (AD) is how to learn one unified and generalizable model to adapt to multi-class especially cross-class settings: the model is trained with normal samples from seen classes with the objective to detect anomalies from both seen and unseen classes. In this work, we propose a novel Proposal Masked Anomaly Detection (PMAD) approach for such challenging multi- and cross-class anomaly detection. The proposed PMAD can be adapted to seen and unseen classes by two key designs: MAE-based patch-level reconstruction and prototype-guided proposal masking. First, motivated by MAE (Masked AutoEncoder), we develop a patch-level reconstruction model rather than the image-level reconstruction adopted in most AD methods for this reason: the masked patches in unseen classes can be reconstructed well by using the visible patches and the adaptive reconstruction capability of MAE. Moreover, we improve MAE by ViT encoder-decoder architecture, combinational masking, and visual tokens as reconstruction objectives to make it more suitable for anomaly detection. Second, we develop a two-stage anomaly detection manner during inference. In the proposal masking stage, the prototype-guided proposal masking module is utilized to generate proposals for suspicious anomalies as much as possible, then masked patches can be generated from the proposal regions. By masking most likely anomalous patches, the “shortcut reconstruction” issue (i.e., anomalous regions can be well reconstructed) can be mostly avoided. In the reconstruction stage, these masked patches are then reconstructed by the trained patch-level reconstruction model to determine if they are anomalies. Extensive experiments show that the proposed PMAD can outperform current state-of-the-art models significantly under the multi- and especially cross-class settings. Code will be publicly available at https://github.com/xcyao00/PMAD.

Most anomaly detection (AD) models are learned using only normal samples in an unsupervised way, which may result in ambiguous decision boundary and insufficient discriminability. In fact, a few anomaly samples are often available in real-world applications, the valuable knowledge of known anomalies should also be effectively exploited. However, utilizing a few known anomalies during training may cause another issue that the model may be biased by those known anomalies and fail to generalize to unseen anomalies. In this paper, we tackle supervised anomaly detection, i.e., we learn AD models using a few available anomalies with the objective to detect both the seen and unseen anomalies. We propose a novel explicit boundary guided semi-push-pull contrastive learning mechanism, which can enhance model's discriminability while mitigating the bias issue. Our approach is based on two core designs: First, we find an explicit and compact separating boundary as the guidance for further feature learning. As the boundary only relies on the normal feature distribution, the bias problem caused by a few known anomalies can be alleviated. Second, a boundary guided semi-push-pull loss is developed to only pull the normal features together while pushing the abnormal features apart from the separating boundary beyond a certain margin region. In this way, our model can form a more explicit and discriminative decision boundary to distinguish known and also unseen anomalies from normal samples more effectively. Code will be available at https://github.com/xcyao00/BGAD.

Humans recognize anomalies through two aspects: larger patch-wise representation discrepancies and weaker patch-to-normal-patch correlations. However, the previous AD methods didn’t sufficiently combine the two complementary aspects to design AD models. To this end, we find that Transformer can ideally satisfy the two aspects as its great power in the unified modeling of patch-wise representations and patch-to-patch correlations. In this paper, we propose a novel AD framework: FOcus-the-Discrepancy (FOD), which can simultaneously spot the patch-wise, intra- and inter-discrepancies of anomalies. The major characteristic of our method is that we renovate the self-attention maps in transformers to Intra-Inter-Correlation (I2Correlation). The I2Correlation contains a two-branch structure to first explicitly establish intra-and inter-image correlations, and then fuses the features of two-branch to spotlight the abnormal patterns. To learn the intra- and inter-correlations adaptively, we propose the RBF-kernel-based target-correlations as learning targets for self-supervised learning. Besides, we introduce an entropy constraint strategy to solve the mode collapse issue in optimization and further amplify the normal-abnormal distinguishability. Extensive experiments on three unsupervised real-world AD benchmarks show the superior performance of our approach. Code will be available at https://github.com/xcyao00/FOD.

Most anomaly detection models are often sensitive to unavoidable disturbance or non-defective "visual defects", and such near abnormal samples are easily identified as anomalies, resulting in a high false detection rate. To this end, we propose a novel multi-scale Cross-image Knowledge Transfer anomaly detection model, namely CKT. Different from most existing intra-image distillation methods, our model transfers both the intra-image knowledge of the normal image and the inter-image knowledge of the normal image and the near-anomaly prototype, to assist the model to learn more robust normal patterns. Furthermore, we develop a cross-image attention module for explicitly enhancing the near-abnormal pattern learning during the distillation procedure, to alleviate the problem of high false detection rate induced by near-abnormal instances. Extensive experiments on texture datasets, such as KSDD2, MT, AITEX, and the textural subset of Mvtec-AD, show that the proposed CKT model can outperform most of the current unsupervised anomaly detection methods. Compared with the existing distillation based anomaly detection frameworks, our work can get significant gains with a margin of 2%.