Papers
Topics
Authors
Recent
2000 character limit reached

Revisiting Deep Feature Reconstruction for Logical and Structural Industrial Anomaly Detection (2410.16255v1)

Published 21 Oct 2024 in cs.CV and cs.LG

Abstract: Industrial anomaly detection is crucial for quality control and predictive maintenance, but it presents challenges due to limited training data, diverse anomaly types, and external factors that alter object appearances. Existing methods commonly detect structural anomalies, such as dents and scratches, by leveraging multi-scale features from image patches extracted through deep pre-trained networks. However, significant memory and computational demands often limit their practical application. Additionally, detecting logical anomalies-such as images with missing or excess elements-requires an understanding of spatial relationships that traditional patch-based methods fail to capture. In this work, we address these limitations by focusing on Deep Feature Reconstruction (DFR), a memory- and compute-efficient approach for detecting structural anomalies. We further enhance DFR into a unified framework, called ULSAD, which is capable of detecting both structural and logical anomalies. Specifically, we refine the DFR training objective to improve performance in structural anomaly detection, while introducing an attention-based loss mechanism using a global autoencoder-like network to handle logical anomaly detection. Our empirical evaluation across five benchmark datasets demonstrates the performance of ULSAD in detecting and localizing both structural and logical anomalies, outperforming eight state-of-the-art methods. An extensive ablation study further highlights the contribution of each component to the overall performance improvement. Our code is available at https://github.com/sukanyapatra1997/ULSAD-2024.git

Definition Search Book Streamline Icon: https://streamlinehq.com
References (55)
  1. On the surprising behavior of distance metrics in high dimensional space. In Database Theory—ICDT 2001: 8th International Conference London, UK, January 4–6, 2001 Proceedings 8, pp.  420–434. Springer, 2001.
  2. A survey of anomaly detection techniques in financial domain. Future Generation Computer Systems, 55:278–288, 2 2016. ISSN 0167-739X. doi: 10.1016/J.FUTURE.2015.01.001.
  3. Anomalib: A Deep Learning Library for Anomaly Detection. In 2022 IEEE International Conference on Image Processing (ICIP), pp.  1706–1710. IEEE, 10 2022. ISBN 978-1-6654-9620-9. doi: 10.1109/ICIP46576.2022.9897283. URL https://ieeexplore.ieee.org/document/9897283/.
  4. Self-supervised learning from images with a joint-embedding predictive architecture. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.  15619–15629, 2023.
  5. EfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latencies. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pp.  128–138, 1 2024. URL https://arxiv.org/abs/2303.14535v2.
  6. MVTEC ad-A comprehensive real-world dataset for unsupervised anomaly detection. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp.  9584–9592, 2019. ISSN 10636919. doi: 10.1109/CVPR.2019.00982.
  7. Uninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.  4183–4192, 2020. URL www.mvtec.com.
  8. Beyond Dents and Scratches: Logical Constraints in Unsupervised Anomaly Detection and Localization. International Journal of Computer Vision, 130(4):947–969, 2022. ISSN 1573-1405. doi: 10.1007/s11263-022-01578-9. URL https://doi.org/10.1007/s11263-022-01578-9.
  9. Anomaly detection. ACM Computing Surveys, 41(3):1–58, 7 2009. ISSN 0360-0300. doi: 10.1145/1541880.1541882. URL https://dl.acm.org/doi/10.1145/1541880.1541882.
  10. Explainable anomaly detection framework for predictive maintenance in manufacturing systems. Applied Soft Computing, 125:109147, 8 2022. ISSN 1568-4946. doi: 10.1016/J.ASOC.2022.109147.
  11. PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localization. In International Conference on Pattern Recognition, volume 12664 LNCS, pp.  475–489. Springer Science and Business Media Deutschland GmbH, 11 2021. ISBN 9783030687984. doi: 10.48550/arxiv.2011.08785. URL https://arxiv.org/abs/2011.08785v1.
  12. Anomaly Detection via Reverse Distillation From One-Class Embedding. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.  9737–9746, 2022.
  13. Generating images with perceptual similarity metrics based on deep networks. Advances in neural information processing systems, 29, 2016.
  14. An image is worth 16x16 words: Transformers for image recognition at scale. In International Conference on Learning Representations, 2021. URL https://openreview.net/forum?id=YicbFdNTTy.
  15. Deep learning for medical anomaly detection – a survey. ACM Comput. Surv., 54(7), jul 2021. ISSN 0360-0300. doi: 10.1145/3464423. URL https://doi.org/10.1145/3464423.
  16. Towards Total Online Unsupervised Anomaly Detection and Localization in Industrial Vision, 5 2023. URL https://arxiv.org/abs/2305.15652v1.
  17. Memorizing Normality to Detect Anomaly: Memory-augmented Deep Autoencoder for Unsupervised Anomaly Detection. Proceedings of the IEEE International Conference on Computer Vision, 2019-October:1705–1714, 4 2019. ISSN 15505499. doi: 10.48550/arxiv.1904.02639. URL https://arxiv.org/abs/1904.02639v2.
  18. CFLOW-AD: Real-Time Unsupervised Anomaly Detection with Localization via Conditional Normalizing Flows. Proceedings - 2022 IEEE/CVF Winter Conference on Applications of Computer Vision, WACV 2022, pp.  1819–1828, 7 2021. doi: 10.1109/WACV51458.2022.00188. URL https://arxiv.org/abs/2107.12571v1.
  19. Financial Fraud: A Review of Anomaly Detection Techniques and Recent Advances. Expert Systems with Applications, 193:116429, 5 2022. ISSN 0957-4174. doi: 10.1016/J.ESWA.2021.116429.
  20. Deep learning-based defect detection of metal parts: Evaluating current methods in complex conditions. International Congress on Ultra Modern Telecommunications and Control Systems and Workshops, 2021-October:66–71, 2021. ISSN 2157023X. doi: 10.1109/ICUMT54235.2021.9631567.
  21. SoftPatch: Unsupervised Anomaly Detection with Noisy Data. Advances in Neural Information Processing Systems, 35:15433–15445, 2022. URL https://github.com/TencentYoutuResearch/AnomalyDetection-SoftPatch.
  22. Robust Subspace Recovery Layer for Unsupervised Anomaly Detection. arXiv preprint, 3 2019. doi: 10.48550/arxiv.1904.00152. URL https://arxiv.org/abs/1904.00152v2.
  23. CFA: Coupled-hypersphere-based Feature Adaptation for Target-Oriented Anomaly Localization. IEEE Access, 10:78446–78454, 6 2022. ISSN 21693536. doi: 10.1109/ACCESS.2022.3193699. URL https://arxiv.org/abs/2206.04325v1.
  24. Deep industrial image anomaly detection: A survey. Machine Intelligence Research, 21(1):104–135, 2024.
  25. VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization. IEEE International Symposium on Industrial Electronics, 2021-June, 4 2021. doi: 10.1109/ISIE45552.2021.9576231. URL http://arxiv.org/abs/2104.10036http://dx.doi.org/10.1109/ISIE45552.2021.9576231.
  26. Anomaly Detection With Multiple-Hypotheses Predictions. In Kamalika Chaudhuri and Ruslan Salakhutdinov (eds.), Proceedings of the 36th International Conference on Machine Learning, volume 97 of Proceedings of Machine Learning Research, pp.  4800–4809. PMLR, 7 2019. URL https://proceedings.mlr.press/v97/nguyen19b.html.
  27. Deep Learning for Anomaly Detection: A Review. ACM Computing Surveys, 54(2), 7 2020. doi: 10.1145/3439950. URL http://arxiv.org/abs/2007.02500http://dx.doi.org/10.1145/3439950.
  28. PyTorch: An Imperative Style, High-Performance Deep Learning Library. Advances in Neural Information Processing Systems, 32, 2019.
  29. OCGAN: One-class novelty detection using gans with constrained latent representations. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2019-June:2893–2901, 6 2019. ISSN 10636919. doi: 10.1109/CVPR.2019.00301.
  30. Transferring the Contamination Factor between Anomaly Detection Domains by Shape Similarity. Proceedings of the AAAI Conference on Artificial Intelligence, 36(4):4128–4136, 6 2022. ISSN 2374-3468. doi: 10.1609/AAAI.V36I4.20331. URL https://ojs.aaai.org/index.php/AAAI/article/view/20331.
  31. Estimating the Contamination Factor’s Distribution in Unsupervised Anomaly Detection. In Proceedings of the 40th International Conference on Machine Learning, pp.  27668–27679. PMLR, 7 2023. URL https://proceedings.mlr.press/v202/perini23a.html.
  32. Anomaly Detection Requires Better Representations. In European Conference on Computer Vision, pp.  56–68, 10 2022. ISBN 9783031250682. doi: 10.1007/978-3-031-25069-9{\_}4. URL https://arxiv.org/abs/2210.10773v1.
  33. Towards Total Recall in Industrial Anomaly Detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.  14318–14328, 2022. doi: 10.48550/arxiv.2106.08265. URL https://arxiv.org/abs/2106.08265v2.
  34. Deep One-Class Classification. In International Conference on Machine Learning, pp.  4393–4402. PMLR, 2018. URL https://proceedings.mlr.press/v80/ruff18a.html.
  35. A Unifying Review of Deep and Shallow Anomaly Detection. Proceedings of the IEEE, 109(5):756–795, 5 2021. ISSN 15582256. doi: 10.1109/JPROC.2021.3052449.
  36. ImageNet Large Scale Visual Recognition Challenge. International Journal of Computer Vision, 115(3):211–252, 12 2015. ISSN 15731405. doi: 10.1007/S11263-015-0816-Y/FIGURES/16. URL https://link.springer.com/article/10.1007/s11263-015-0816-y.
  37. Adversarially Learned One-Class Classifier for Novelty Detection. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp.  3379–3388, 12 2018. ISSN 10636919. doi: 10.1109/CVPR.2018.00356.
  38. Multiresolution knowledge distillation for anomaly detection. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp.  14897–14907, 11 2021. ISSN 10636919. doi: 10.1109/CVPR46437.2021.01466. URL https://arxiv.org/abs/2011.11108v1.
  39. A unified survey on anomaly, novelty, open-set, and out of-distribution detection: Solutions and future challenges. Transactions on Machine Learning Research, 2022. ISSN 2835-8856. URL https://openreview.net/forum?id=aRtjVZvbpK.
  40. Deep learning based automatic defect identification of photovoltaic module using electroluminescence images. Solar Energy, 201:453–460, 5 2020. ISSN 0038-092X. doi: 10.1016/J.SOLENER.2020.03.049.
  41. Outlier sums for differential gene expression analysis. Biostatistics, 8(1):2–8, 1 2007. ISSN 1465-4644. doi: 10.1093/BIOSTATISTICS/KXL005. URL https://academic.oup.com/biostatistics/article/8/1/2/252461.
  42. Similarity-preserving knowledge distillation. In Proceedings of the IEEE/CVF international conference on computer vision, pp.  1365–1374, 2019.
  43. Effective End-to-end Unsupervised Outlier Detection via Inlier Priority of Discriminative Network. Advances in Neural Information Processing Systems, 32, 2019.
  44. Learning Semantic Context from Normal Samples for Unsupervised Anomaly Detection. Proceedings of the AAAI Conference on Artificial Intelligence, 35(4):3110–3118, 5 2021. URL https://ojs.aaai.org/index.php/AAAI/article/view/16420.
  45. DFR: Deep Feature Reconstruction for Unsupervised Anomaly Segmentation. Neurocomputing, 424:9–22, 12 2020. doi: 10.1016/j.neucom.2020.11.018. URL http://arxiv.org/abs/2012.07122http://dx.doi.org/10.1016/j.neucom.2020.11.018.
  46. Self-supervise, Refine, Repeat: Improving Unsupervised Anomaly Detection. Transactions on Machine Learning Research (TMLR), 2022. URL https://arxiv.org/abs/2106.06115v2.
  47. SPADE: Semi-supervised Anomaly Detection under Distribution Mismatch. Transactions on Machine Learning Research, 11 2023. URL http://arxiv.org/abs/2212.00173.
  48. FastFlow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flows. arXiv preprint arXiv:2111.07677, 11 2021. doi: 10.48550/arxiv.2111.07677. URL https://arxiv.org/abs/2111.07677v2.
  49. Wide Residual Networks. British Machine Vision Conference 2016, BMVC 2016, 2016-September:1–87, 5 2016. doi: 10.5244/C.30.87. URL https://arxiv.org/abs/1605.07146v4.
  50. Paying more attention to attention: Improving the performance of convolutional neural networks via attention transfer. In International Conference on Learning Representations, 2017. URL https://openreview.net/forum?id=Sks9_ajex.
  51. DRÆM - A discriminatively trained reconstruction embedding for surface anomaly detection. Proceedings of the IEEE International Conference on Computer Vision, pp.  8310–8319, 8 2021. ISSN 15505499. doi: 10.1109/ICCV48922.2021.00822. URL https://arxiv.org/abs/2108.07610v2.
  52. Efficient GAN-Based Anomaly Detection. arXiv preprint, 2 2018. doi: 10.48550/arxiv.1802.06222. URL https://arxiv.org/abs/1802.06222v2.
  53. Contextual Affinity Distillation for Image Anomaly Detection. arXiv preprint arXiv:2307.03101, 7 2023. URL https://arxiv.org/abs/2307.03101v1.
  54. Encoding Structure-Texture Relation with P-Net for Anomaly Detection in Retinal Images. In European conference on computer vision, volume 12365 LNCS, pp.  360–377. Springer, 2020. doi: 10.1007/978-3-030-58565-5{\_}22. URL https://link.springer.com/10.1007/978-3-030-58565-5_22.
  55. Spot-the-difference self-supervised pre-training for anomaly detection and segmentation. In European Conference on Computer Vision, pp.  392–408. Springer, 2022.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now

Whiteboard

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Generate Now

Continue Learning

We haven't generated follow-up questions for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Generate Now

Collections

Sign up for free to add this paper to one or more collections.

User Circle Single Streamline Icon: https://streamlinehq.com Sign Up