Toward Long-Tailed Online Anomaly Detection through Class-Agnostic Concepts
Not on the map. Why?
Abstract (arXiv)
Anomaly detection (AD) identifies the defect regions of a given image. Recent works have studied AD, focusing on learning AD without abnormal images, with long-tailed distributed training data, and using a unified model for all classes. In addition, online AD learning has also been explored. In this work, we expand in both directions to a realistic setting by considering the novel task of long-tailed online AD (LTOAD). We first identified that the offline state-of-the-art LTAD methods cannot be directly applied to the online setting. Specifically, LTAD is class-aware, requiring class labels that are not available in the online setting. To address this challenge, we propose a class-agnostic framework for LTAD and then adapt it to our online learning setting. Our method outperforms the SOTA baselines in most offline LTAD settings, including both the industrial manufacturing and the medical domain. In particular, we observe +4.63% image-AUROC on MVTec even compared to methods that have access to class labels and the number of classes. In the most challenging long-tailed online setting, we achieve +0.53% image-AUROC compared to baselines. Our LTOAD benchmark is released here: https://doi.org/10.5281/zenodo.16283852 .
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 76 of its 98 references that are in the map. First posted on arXiv 22 Jul 2025.
Builds on
76 of 98 references.
References · 76
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
- 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
- 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,442
- 2017Neural Discrete Representation Learningcited by 4,712
- 2021Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervisioncited by 3,864
- 2015Variational Inference with Normalizing Flowscited by 2,170
- 2023A Comprehensive Survey of Continual Learning: Theory, Method and Applicationcited by 872
- 2021Towards Total Recall in Industrial Anomaly Detectioncited by 776
- 2021Scaling Vision with Sparse Mixture of Expertscited by 747
- 2020PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localizationcited by 594
- 2021CutPaste: Self-Supervised Learning for Anomaly Detection and Localizationcited by 531
38 more
- 2021The RSNA-ASNR-MICCAI BraTS 2021 Benchmark on Brain Tumor Segmentation and Radiogenomic Classificationcited by 518
- 2021DRAEM -- A discriminatively trained reconstruction embedding for surface anomaly detectioncited by 514
- 2019The Liver Tumor Segmentation Benchmark (LiTS)cited by 509
- 2022Anomaly Detection via Reverse Distillation from One-Class Embeddingcited by 421
- 2022SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentationcited by 400
- 2023WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentationcited by 315
- 2020Sub-Image Anomaly Detection with Deep Pyramid Correspondencescited by 300
- 2020Multiresolution Knowledge Distillation for Anomaly Detectioncited by 285
- 2023AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly Detectioncited by 271
- 2022A Unified Model for Multi-class Anomaly Detectioncited by 268
- 2021FastFlow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flowscited by 227
- 2020Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flowscited by 212
- 2023OneLLM: One Framework to Align All Modalities with Languagecited by 187
- 2023AnomalyGPT: Detecting Industrial Anomalies Using Large Vision-Language Modelscited by 171
- 2022DeSTSeg: Segmentation Guided Denoising Student-Teacher for Anomaly Detectioncited by 154
- 2023Meta-Transformer: A Unified Framework for Multimodal Learningcited by 136
- 2022Registration based Few-Shot Anomaly Detectioncited by 133
- 2024AdaCLIP: Adapting CLIP with Hybrid Learnable Prompts for Zero-Shot Anomaly Detectioncited by 127
- 2024MambaAD: Exploring State Space Models for Multi-class Unsupervised Anomaly Detectioncited by 113
- 2024PromptAD: Learning Prompts with only Normal Samples for Few-Shot Anomaly Detectioncited by 98
- 2024Toward Generalist Anomaly Detection via In-context Residual Learning with Few-shot Sample Promptscited by 89
- 2023BMAD: Benchmarks for Medical Anomaly Detectioncited by 82
- 2021Defect-GAN: High-Fidelity Defect Synthesis for Automated Defect Inspectioncited by 77
- 2024A Unified Anomaly Synthesis Strategy with Gradient Ascent for Industrial Anomaly Detection and Localizationcited by 74
- 2024A Survey on Visual Anomaly Detection: Challenge, Approach, and Prospectcited by 69
- 2024Adapting Visual-Language Models for Generalizable Anomaly Detection in Medical Imagescited by 69
- 2021Online Continual Learning on Class Incremental Blurry Task Configuration with Anytime Inferencecited by 67
- 2023Hierarchical Vector Quantized Transformer for Multi-class Unsupervised Anomaly Detectioncited by 66
- 2024MuSc: Zero-Shot Industrial Anomaly Classification and Segmentation with Mutual Scoring of the Unlabeled Imagescited by 52
- 2024FiLo: Zero-Shot Anomaly Detection by Fine-Grained Description and High-Quality Localizationcited by 51
- 2024VCP-CLIP: A visual context prompting model for zero-shot anomaly segmentationcited by 50
- 2020Continual Learning for Anomaly Detection in Surveillance Videoscited by 49
- 2024ClipSAM: CLIP and SAM Collaboration for Zero-Shot Anomaly Segmentationcited by 49
- 2023Real-Time Evaluation in Online Continual Learning: A New Hopecited by 47
- 2023Exploring Plain ViT Reconstruction for Multi-class Unsupervised Anomaly Detectioncited by 47
- 2020Anomaly detection through latent space restoration using vector-quantized variational autoencoderscited by 45
- 2023TransFusion -- A Transparency-Based Diffusion Model for Anomaly Detectioncited by 43
- 2024GLAD: Towards Better Reconstruction with Global and Local Adaptive Diffusion Models for Unsupervised Anomaly Detectioncited by 42
50 most cited of 76 All 76 (JSON)
Built on by
Citing papers per year: 2025: 1.
1 paper.