InCLAD: A Continual Learning Benchmark for Industrial Visual Anomaly Detection

Łukasz Marcjan; Dorota Wilk-Kołodziejczyk; Roberto Corizzo

cs.LGOct 10, 2026

InCLAD: A Continual Learning Benchmark for Industrial Visual Anomaly Detection

Authors: Łukasz Marcjan, Dorota Wilk-Kołodziejczyk, Roberto Corizzo

Organizations: AGH University of Krakow, Krakow, Poland · American University, Washington, DC, USA

Abstract

Industrial anomaly detection benchmarks have advanced rapidly, yet they remain largely static and therefore underrepresent deployment settings in which inspection systems must be extended from one industrial component to another over time. In our setting, continual industrial anomaly detection is framed as a task-incremental problem: each component defines a task, while the learning objective remains semi-supervised. The goal extends beyond conventional anomaly detection, since models should preserve their performance on all components as they are sequentially exposed to new components. This framing is especially relevant because many existing industrial pipelines either assume fully supervised defect labels or study only isolated datasets, leaving open the question of how semi-supervised anomaly detectors behave across a sequence of heterogeneous components. We address this gap by organizing the benchmark as a sequence of related component-level tasks and by defining multiple scenarios, including random, easy-to-hard, and hard-to-easy. In addition to single-dataset continual scenarios, we propose challenging scenarios spanning across multiple datasets. We further evaluate continual performance from two complementary perspectives: image-level anomaly detection and pixel-level anomaly localization. By recasting industrial anomaly detection as a continual benchmark rather than a static train/test problem, our work establishes a concrete path toward studying knowledge retention, adaptation, and generalization in evolving industrial environments.

Figures & tables

Panel of Figure 1 — Figure 1: Our InCLAD benchmark at a glance. Tasks are drawn from five industrial datasets plus a multi-dataset stream, then arranged as curriculum or random continual scenario orderings. We evaluate 9 methods under non-continual (NCL) and continual (CL) protocols from both the image-level and pixel-level perspectives using ROC-AUC and BWT.

Panel of Figure 2 — Figure 2: Single-task ROC-AUC heatmaps used to construct the benchmark scenarios. For each dataset, rows are ordered by the consensus curriculum induced by the average rank across PatchCore, PaDiM, and CFA, from easier concepts at the top to harder concepts at the bottom. The heatmap entries are Single-Task Expert ROC-AUC values, and the final column reports their average.

Scenario	# Tasks	# Data samples	# Anomalies (test)
BTAD	3	2,540	39.14%
DAGM	10	15,054	13.09%
MPDD	6	1,346	61.57%
MVTec AD	15	5,354	72.93%
VisA	12	10,821	55.50%
MultiDataset	27	20,853	29.38%

Table 1: Overview of the proposed visual anomaly detection scenarios. Each scenario comes in two curriculum orderings (easy-to-hard, hard-to-easy). Single-dataset scenarios are also provided with one additional random ordering, leading to 17 scenarios in total.

Dataset	Ordering	Metric	Image-level		Pixel-level
Dataset	Ordering	Metric	NCL	CL	NCL	CL
MVTec	Easy-to-Hard	ROC $\uparrow$	$0.583\,{\scriptstyle\pm\,0.095}$	$0.784\,{\scriptstyle\pm\,0.146}$	$0.629\,{\scriptstyle\pm\,0.128}$	$0.842\,{\scriptstyle\pm\,0.105}$
	Easy-to-Hard	BWT $\uparrow$	$-0.039\,{\scriptstyle\pm\,0.018}$	$-0.017\,{\scriptstyle\pm\,0.015}$	$-0.035\,{\scriptstyle\pm\,0.008}$	$-0.010\,{\scriptstyle\pm\,0.008}$
	Hard-to-Easy	ROC $\uparrow$	$0.557\,{\scriptstyle\pm\,0.062}$	$0.710\,{\scriptstyle\pm\,0.117}$	$0.623\,{\scriptstyle\pm\,0.125}$	$0.839\,{\scriptstyle\pm\,0.103}$
	Hard-to-Easy	BWT $\uparrow$	$-0.042\,{\scriptstyle\pm\,0.016}$	$-0.017\,{\scriptstyle\pm\,0.015}$	$-0.042\,{\scriptstyle\pm\,0.010}$	$-0.010\,{\scriptstyle\pm\,0.006}$
	Random	ROC $\uparrow$	$0.593\,{\scriptstyle\pm\,0.037}$	$0.766\,{\scriptstyle\pm\,0.124}$	$0.679\,{\scriptstyle\pm\,0.080}$	$0.858\,{\scriptstyle\pm\,0.090}$

Table 2: ROC-AUC and BWT averaged across methods and ResNet18/WideResNet backbones. NCL denotes the naive baseline, while CL denotes the continual variant.

Appendix figures & tables37 assets

Supplementary material from the paper’s appendix.

Appendix

Benchmark	Industrial	Continual	Normal-only	Image	Pixel	Curricula	Multi-dataset	BWT
	domain	/ task-inc.	training	eval.	eval.	orderings	stream	metrics
MVTec AD [ Bergmann et al., 2019 ]	Yes	No	Yes	Yes	Yes	No	No	No
VisA [ Zou et al., 2022 ]	Yes	No	Yes	Yes	Yes	No	No	No
ADBench [ Han et al., 2022 ]	Partial	No	Partial	No	No	No	Yes	No
CableInspect-AD [ Arodi et al., 2024 ]	Yes	No	Partial	Yes	Yes	No	No	No
ReinAD [ Wang et al., 2025 ]	Yes	No	Yes	Yes	Yes	No	No	No

Appendix

Table 3: Comparison with representative anomaly-detection benchmarks. “Partial” indicates that the property is covered only for some datasets, tasks, or evaluation settings.

Dataset	Model	Time	Naive	Continual
MVTec	CFA	Train	6m 32s ( $\pm$ 1m 42s)	15m 42s ( $\pm$ 10m 57s)
	CFA	Eval	9m 35s ( $\pm$ 1m 40s)	52m 7s ( $\pm$ 31m 50s)
	PaDiM	Train	6m 19s ( $\pm$ 1m 39s)	7m 9s ( $\pm$ 1m 18s)
	PaDiM	Eval	30m 28s ( $\pm$ 3m 22s)	65m 24s ( $\pm$ 3m 37s)
	PatchCore	Train	3m 26s ( $\pm$ 12.6s)	4m 17s ( $\pm$ 10.2s)
	PatchCore	Eval	34m 27s ( $\pm$ 3m 58s)	49m 49s ( $\pm$ 1m 42s)

Appendix

Table 4: Execution times (train and eval separately, mean

\pm

population std across available orderings) for WideResNet-50-2-based models. Continual denotes CL for CFA, PaDiM, and PatchCore, and replay for the remaining models. When image and pixel results were produced as separate runs, their times are summed to reflect the total wall-clock cost of obtaining both modalities.

Dataset	Model	Time	Naive	Continual
MVTec	GANomaly (default)	Train	4m 59s ( $\pm$ 0.7s)	5m 56s ( $\pm$ 6.1s)
	GANomaly (default)	Eval	9m 38s ( $\pm$ 8.4s)	9m 44s ( $\pm$ 7.1s)
	EfficientAD (small)	Train	–	80m 15s ( $\pm$ 1m 6s)
	EfficientAD (small)	Eval	–	11m 55s ( $\pm$ 1.4s)
VisA	GANomaly (default)	Train	11m 32s ( $\pm$ 10.5s)	13m 23s ( $\pm$ 18.4s)
	GANomaly (default)	Eval	10m 44s ( $\pm$ 35.1s)	10m 35s ( $\pm$ 41.3s)

Appendix

Table 5: Execution times (train and eval separately, mean

\pm

population std across available orderings) for models without the two-backbone setup: GANomaly (default) and EfficientAD (small). When image and pixel results were produced as separate runs, their times are summed to reflect the total wall-clock cost of obtaining both modalities.

Training
Hyperparameter	CFA	PaDiM	PatchCore
Backbone	resnet18 / wide_resnet50_2	resnet18 / wide_resnet50_2	resnet18 / wide_resnet50_2
Pretrained	✓	✓	✓
Batch size	32	32	32
Epochs	50	—	—
Learning rate	$10^{-3}$	—	—

Appendix

Table 6: Effective hyperparameters of the CL-specific visual anomaly detection models used in our experiments. The runner passes epochs=50 to all CL jobs, but this setting is effective only for CFA.

Training
Hyperparameter	FastFlow	GANomaly	PaSTe	RD4AD	STFPM	EfficientAD
Backbone	resnet18 / wide_resnet50_2	—	resnet18 / wide_resnet50_2	resnet18 / wide_resnet50_2	resnet18 / wide_resnet50_2	—
Pretrained	✓	—	✓	✓	✓	—
Batch size	8	16	32	16	32	1
Epochs	200	15	100	200	100	50
Learning rate	$10^{-3}$	$2\cdot 10^{-4}$	0.4	$5\cdot 10^{-3}$	0.4	$10^{-4}$

Appendix

Table 7: Effective hyperparameters of the non-CL visual anomaly detection models used in our experiments. When two backbones are listed, separate runs were executed for both settings.

Dataset	PatchCore vs PaDiM	PatchCore vs CFA	PaDiM vs CFA
BTAD	0.500	1.000	0.500
DAGM	0.855	0.455	0.648
MPDD	-0.029	0.200	-0.657
MVTec AD	0.382	0.693	0.654
VisA	0.622	0.797	0.748

Appendix

Table 8: Pairwise Spearman rank correlations between the three Single-Task Expert models for each dataset. Higher values indicate stronger agreement on concept difficulty, while low or negative values indicate that the relative difficulty of concepts depends more strongly on the chosen detector.

Rank	Concept	Avg. ROC-AUC	Avg. Rank	PatchCore	PaDiM	CFA
1	02	0.7947	1.00	0.7923	0.7955	0.7963
2	01	0.9679	2.33	0.9456	0.9874	0.9708
3	03	0.9903	2.67	0.9999	0.9819	0.9890

Appendix

Table 9: BTAD hard-to-easy concept ordering derived from average rank across PatchCore, PaDiM, and CFA. The easy-to-hard scenario is the reverse of this table.

Rank	Concept	Avg. ROC-AUC	Avg. Rank	PatchCore	PaDiM	CFA
1	Class8	0.6541	1.00	0.6729	0.7278	0.5616
2	Class5	0.8210	3.00	0.8164	0.8160	0.8308
3	Class2	0.8594	3.67	0.8920	0.9027	0.7835
4	Class1	0.8380	3.67	0.8561	0.7395	0.9183
5	Class3	0.8625	4.67	0.9776	0.9274	0.6825
6	Class7	0.9565	7.33	0.9990	0.9310	0.9394

Appendix

Table 10: DAGM hard-to-easy concept ordering derived from average rank across PatchCore, PaDiM, and CFA.

Rank	Concept	Avg. ROC-AUC	Avg. Rank	PatchCore	PaDiM	CFA
1	bracket_black	0.6791	2.33	0.8471	0.6410	0.5492
2	tubes	0.7435	3.00	0.8170	0.5738	0.8397
3	bracket_white	0.7537	3.00	0.7311	0.8533	0.6767
4	metal_plate	0.8106	3.67	0.9420	0.4902	0.9995
5	bracket_brown	0.8286	4.33	0.9480	0.7624	0.7753
6	connector	0.8643	4.67	1.0000	0.8214	0.7714

Appendix

Table 11: MPDD hard-to-easy concept ordering derived from average rank across PatchCore, PaDiM, and CFA.

Rank	Concept	Avg. ROC-AUC	Avg. Rank	PatchCore	PaDiM	CFA
1	pill	0.8300	2.00	0.8161	0.7700	0.9037
2	screw	0.7930	3.00	0.9664	0.7284	0.6842
3	capsule	0.8738	3.67	0.9334	0.8337	0.8544
4	grid	0.9176	4.67	0.8956	0.9181	0.9390
5	zipper	0.9079	5.33	0.9362	0.8277	0.9598
6	transistor	0.9193	7.00	0.9729	0.8129	0.9721

Appendix

Table 12: MVTec AD hard-to-easy concept ordering derived from average rank across PatchCore, PaDiM, and CFA.

Rank	Concept	Avg. ROC-AUC	Avg. Rank	PatchCore	PaDiM	CFA
1	capsules	0.6493	1.33	0.7465	0.6175	0.5840
2	macaroni2	0.6407	2.00	0.6722	0.6567	0.5931
3	macaroni1	0.8148	4.00	0.8965	0.7791	0.7687
4	pcb3	0.7765	4.33	0.9196	0.6352	0.7746
5	pcb2	0.8363	6.00	0.9055	0.8094	0.7941
6	candle	0.8623	6.00	0.8770	0.7839	0.9259

Appendix

Table 13: VisA hard-to-easy concept ordering derived from average rank across PatchCore, PaDiM, and CFA.

Figure 3: BTAD , ResNet-18 : concept transition matrices across three orderings (Hard

\to

Easy, Easy

\to

Hard, Random): NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 4: BTAD , Wide ResNet-50-2 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 5: DAGM , ResNet-18 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 6: DAGM , Wide ResNet-50-2 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 7: MPDD , ResNet-18 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 8: MPDD , Wide ResNet-50-2 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 9: MVTec , ResNet-18 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 10: MVTec , Wide ResNet-50-2 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 11: VisA , ResNet-18 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 12: VisA , Wide ResNet-50-2 : concept transition matrices: NCL (top) vs. CL (bottom), image-level and pixel-level.

Figure 13: MultiDataset , ResNet-18, Hard-to-Easy: NCL, image-level concept transition matrix.

Figure 14: MultiDataset , ResNet-18, Hard-to-Easy: CL, image-level concept transition matrix.

Figure 15: MultiDataset , ResNet-18, Hard-to-Easy: NCL, pixel-level concept transition matrix.

Figure 16: MultiDataset , ResNet-18, Hard-to-Easy: CL, pixel-level concept transition matrix.

Figure 17: MultiDataset , ResNet-18, Easy-to-Hard: NCL, pixel-level concept transition matrix.

Figure 18: MultiDataset , ResNet-18, Easy-to-Hard: CL, image-level concept transition matrix.

Figure 19: MultiDataset , ResNet-18, Easy-to-Hard: NCL, image-level concept transition matrix.

Figure 20: MultiDataset , ResNet-18, Easy-to-Hard: CL, pixel-level concept transition matrix.

Figure 21: MultiDataset , Wide ResNet-50-2, Hard-to-Easy: NCL, image-level concept transition matrix.

Figure 22: MultiDataset , Wide ResNet-50-2, Hard-to-Easy: CL, image-level concept transition matrix.

Figure 23: MultiDataset , Wide ResNet-50-2, Hard-to-Easy: NCL, pixel-level concept transition matrix.

Figure 24: MultiDataset , Wide ResNet-50-2, Hard-to-Easy: CL, pixel-level concept transition matrix.

Figure 25: MultiDataset , Wide ResNet-50-2, Easy-to-Hard: NCL, pixel-level concept transition matrix.

Figure 26: MultiDataset , Wide ResNet-50-2, Easy-to-Hard: CL, image-level concept transition matrix.

Figure 27: MultiDataset , Wide ResNet-50-2, Easy-to-Hard: NCL, image-level concept transition matrix.

Figure 28: MultiDataset , Wide ResNet-50-2, Easy-to-Hard: CL, pixel-level concept transition matrix.

Explore similar work

Sep 3, 2026

Neural-Collapse-guided Task-Free Continual Anomaly Detection

Recent years have witnessed growing interest in continual anomaly detection for industrial visual inspection. However, real-world manufacturing environments exhibit unpredictable shifts in data distributions, rendering task-dependent continual learning assumptions impractical. To address this limitation, we formulate industrial anomaly detection as a task-free continual learning problem and propose NC-TFAD, a neural-collapse-inspired, geometry-driven framework for learning from non-stationary data streams without task boundaries. NC-TFAD freezes a pretrained backbone and aligns streaming features to a simplex Equiangular Tight Frame (ETF) prototype space to stabilize representation geometry under non-stationary streams. To satisfy the NC-inspired geometric construction in the absence of real anomalies, we generate synthetic anomaly samples as auxiliary anchors during training. Building on this geometry, we further introduce inter- and intra-class regularization together with a Focal Neural Collapse Contrastive (FNCC) loss to suppress representation drift and improve normal-anomaly separability. Finally, a normal-patch-prototype-guided localization branch constructs calibrated patch-wise deviation maps from normal training samples and fuses them with a weak self-attention prior, producing anomaly heatmaps without pixel-level annotations. Extensive experiments on MVTec AD and VisA show that NC-TFAD consistently outperforms representative task-free continual learning methods adapted from general vision, as well as unified anomaly detection baselines, in both image-level detection and pixel-level localization under the task-free continual learning protocol. These results highlight that geometry-driven modeling offers an effective and robust solution for task-free continual anomaly detection in real-world industrial applications.

Xiaotong Kong, Chaoyang Song, Ziai Zhou +3

Key Laboratory of Measurement and Control of CSE, Ministry of Education, School of Automation, Southeast University, Nanjing 210096, China

Continual Learning Industrial Anomaly Detection Anomaly Localization Anomaly Detection

Ask PDF

May 11, 2026

MMVIAD: Multi-view Multi-task Video Understanding for Industrial Anomaly Detection

Industrial anomaly detection is critical for manufacturing quality control, yet existing datasets mainly focus on static images or sparse views, which do not fully reflect continuous inspection processes in real industrial scenarios. We introduce MMVIAD (Multi-view Multi-task Video Industrial Anomaly Detection), to the best of our knowledge the first continuous multi-view video dataset for industrial anomaly detection and understanding, together with a benchmark for multi-task evaluation. MMVIAD contains object-centric 2-second inspection clips with approximately 120 degrees of camera motion, covering 48 object categories, 14 environments, and 6 structural anomaly types. It supports anomaly detection, defect classification, object classification, and anomaly visible-time localization. Systematic evaluations on MMVIAD show that current commercial and open-source video MLLMs remain far below human performance, especially for fine-grained defect recognition and temporal grounding. To improve transferable anomaly understanding, we further develop a two-stage post-training pipeline where PS-SFT (Perception-Structured Supervised Fine-Tuning) initializes perception-structured reasoning and VISTA-GRPO (Visibility-grounded Industrial Structured Temporal Anomaly Group Relative Policy Optimization) refines the model with semantic-gated defect reward and visibility-aware temporal reward, producing the final model VISTA. On MMVIAD-Unseen, VISTA improves the base model's average score across the four tasks from 45.0 to 57.5, surpassing GPT-5.4. Source code is available at https://github.com/Georgekeepmoving/MMVIAD.

Xiran Zhao, Jing Jin, Yan Bai +6

ShanghaiTech University · Tsinghua University · Meituan Inc. +1

Multi-View Learning Industrial Anomaly Detection Multi-Task Learning Anomaly Localization Multimodal Anomaly Detection Video Anomaly Detection

Ask PDF

Mar 14, 2026

AD-Copilot: A Vision-Language Assistant for Industrial Anomaly Detection via Visual In-context Comparison

Multimodal Large Language Models (MLLMs) have achieved impressive success in natural visual understanding, yet they consistently underperform in industrial anomaly detection (IAD). This is because MLLMs trained mostly on general web data differ significantly from industrial images. Moreover, they encode each image independently and can only compare images in the language space, making them insensitive to subtle visual differences that are key to IAD. To tackle these issues, we present AD-Copilot, an interactive MLLM specialized for IAD via visual in-context comparison. We first design a novel data curation pipeline to mine inspection knowledge from sparsely labeled industrial images and generate precise samples for captioning, VQA, and defect localization, yielding a large-scale multimodal dataset Chat-AD rich in semantic signals for IAD. On this foundation, AD-Copilot incorporates a novel Comparison Encoder that employs cross-attention between paired image features to enhance multi-image fine-grained perception, and is trained with a multi-stage strategy that incorporates domain knowledge and gradually enhances IAD skills. In addition, we introduce MMAD-BBox, an extended benchmark for anomaly localization with bounding-box-based evaluation. The experiments show that AD-Copilot achieves 82.3% accuracy on the MMAD benchmark, outperforming all other models without any data leakage. In the MMAD-BBox test, it achieves a maximum improvement of

3.35\times

over the baseline. AD-Copilot also exhibits excellent generalization of its performance gains across other specialized and general-purpose benchmarks. Remarkably, AD-Copilot surpasses human expert-level performance on several IAD tasks, demonstrating its potential as a reliable assistant for real-world industrial inspection. All datasets and models will be released for the broader benefit of the community.

Xi Jiang, Yue Guo, Jian Li +7

Department of Computer Science and Engineering, Southern University of Science and Technology, Shenzhen, China · Macau University of Science and Technology, Macau, China · Tencent YouTu Lab, Shenzhen and Shanghai, China +3

Industrial Anomaly Detection Industrial Inspection Anomaly Localization In-Context Learning

Ask PDF