Real-Time Anomaly Detection in Packaged Food X-Ray Images Using Supervised Learning
Source abstract
Abstract
Physical contamination of food occurs when it comes into contact with foreign objects. Foreign objects can be introduced to food at any time during food delivery and packaging and can cause serious concerns such as broken teeth or choking. Therefore, a preventive method that can detect and remove foreign objects in advance is required. Several studies have attempted to detect defective products using deep learning networks. Because it is difficult to obtain foreign object-containing food data from industry, most studies on industrial anomaly detection have used unsupervised learning methods. This paper proposes a new method for real-time anomaly detection in packaged food products using a supervised learning network. In this study, a realistic X-ray image training dataset was constructed by augmenting foreign objects with normal product images in a cut-paste manner. Based on the augmented training dataset, we trained YOLOv4, a real-time object detection network, and detected foreign objects in the test data. We evaluated this method on images of pasta, snacks, pistachios, and red beans under the same conditions. The results show that the normal and defective products were classified with an accuracy of at least 94% for all packaged foods. For detecting foreign objects that are typically difficult to detect using the unsupervised learning and traditional methods, the proposed method achieved high-performance real-time anomaly detection. In addition, to eliminate the loss in high-resolution X-ray images, the false positive rate and accuracy could be lowered to 5% with patch-based training and a new post-processing algorithm.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Industrial Vision Systems and Defect Detection · Food Safety and Hygiene · Identification and Quantification in Food
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (64)
Show source-provided references
1. A robust bayesian probabilistic matrix factorization model for collaborative filtering recommender systems based on user anomaly rating behavior detection
2. A novel network anomaly detection method based on data balancing and recursive feature addition
3. Automated fish bone detection using X-ray imaging10.1016/j.jfoodeng.2011.03.007 ↗
4. Davies · Image Processing for the Food Industry · 2000 · 37 · 18510.1142/9789812797636_0011 ↗
5. Graves · Machine Vision for the Inspection of Natural Products · 2003 · 1 · 42110.1007/1-85233-853-9_16 ↗
6. Deep-anomaly: Fully convolutional neural network for fast anomaly detection in crowded scenes10.1016/j.cviu.2018.02.006 ↗
7. Deep semi-supervised anomaly detection
8. Generative probabilistic novelty detection with adversarial autoencoders
9. Novelty detection via blurring
10. CIFAR-10: KNN-based ensemble of classifiers
11. M. Fatahi, “Mnist handwritten digits,” 2014. [Online]. Available: https://www.researchgate.net/profile/Mazdak_Fatahi/publication/273124795_MNIST_handwritten_digits_Description_and_using/links/54f7e82f0cf28d6dec9f6d2d/MNIST-handwritten-digits-Description-and-using.pdf.
12. Deep residual learning for image recognition
13. Going deeper with convolutions
14. EfficientNet: Rethinking model scaling for convolutional neural networks
15. A. Bochkovskiy, C. Y. Wang and H. Y. M. Liao, “YOLOv4: Optimal speed and accuracy of object detection,” arXiv preprint arXiv:2004.10934, 2020.
16. X-ray detection of defects and contaminants in the food industry10.1007/s11694-008-9059-8 ↗
17. Comparison research on X-ray image quality defects detection algorithm in semiconductor manufacturing
18. Novelty detection of foreign objects in food using multi-modal X-ray imaging10.1016/j.foodcont.2016.02.023 ↗
19. G. Pang, C. Shen, L. Cao and A. V. D. Hengel, “Deep learning for anomaly detection: A review,” arXiv preprint arXiv:2007.02500, 2020.
20. R. Chalapathy and S. Chawla, “Deep learning for anomaly detection: A survey,” arXiv preprint arXiv: 1901. 03407, 2019.
21. D. Zimmerer, S. A. A. Kohl, J. Petersen, F. Isensee and K. H. Maier-Hein, “Context-encoding variational autoencoder for unsupervised anomaly detection,” arXiv preprint arXiv: 1812.05941v1, 2018.
22. R. Chalapathy, A. K. Menon and S. Chawla, “Anomaly detection using one-class neural networks,” arXiv preprint arXiv: 1802. 06360, 2018.
23. Deep one-class classification
24. ImageNet: A large-scale hierarchical image database
25. Defect detection in SEM images of nanofibrous materials10.1109/TII.2016.2641472 ↗
26. Efficient anomaly detection via matrix sketching
27. Baur · Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries · 2019 · 161
28. Auto-encoding variational bayes
29. Goodfellow · Advances in Neural Information Processing Systems · 2014 · 2672
30. Unsupervised anomaly detection with generative adversarial networks to guide marker discovery
31. Clustering and unsupervised anomaly detection with l2 normalized deep auto-encoder representations
32. Improving unsupervised defect segmentation by applying structural similarity to autoencoders
33. GANomaly: Semi-supervised anomaly detection via adversarial training
34. MVTec AD–-a comprehensive real-world dataset for unsupervised anomaly detection
35. U-Net: Convolutional networks for biomedical image segmentation
36. Fully convolutional networks for semantic segmentation
37. SSD: Single shot multi box detector
38. Ren · Advances in Neural Information Processing Systems 28 · 2015 · 91
39. Dai · Advances in Neural Information Processing Systems · 2016 · 379
40. MASK R-CNN
41. You only look once: Unified, real-time object detection
42. Focal loss for dense object detection
43. EfficientDet: Scalable and efficient object detection
44. SpineNet: Learning scale-permuted backbone for recognition and localization
45. Microsoft COCO: Common objects in context
46. Single-shot refinement neural network for object detection
47. YOLO9000: Better, faster, stronger
48. J. Redmon and A. Farhadi, “YOLOv3: An incremental improvement,” arXiv preprint arXiv: 1804. 02767, 2018.
49. DOPING: Generative data augmentation for unsupervised anomaly detection with GAN
50. K. Babaei, Z. Chen and T. Maul, “Data augmentation by autoencoders for unsupervised anomaly detection,” arXiv preprint arXiv: 1808. 07632, 2018.
51. AnoGen: Deep anomaly generator
52. SNIPER: Efficient multi-scale training
53. RICAP: Random image cropping and patching data augmentation for deep CNNs
54. Random erasing data augmentation
55. Modeling visual context is key to augmenting object detection datasets
56. Cut, paste and learn: surprisingly easy synthesis for instance detection
57. Very deep convolutional networks for large-scale image recognition
58. GPipe: Efficient training of giant neural networks using pipeline parallelism
59. IterNet: Retinal image segmentation utilizing structural redundancy in vessel networks
60. Training CNNs with image patches for object localisation10.1049/el.2017.4725 ↗
61. J. Yi and S. Yoon, “Patch SVDD: Patch-level SVDD for anomaly detection and segmentation,” arXiv preprint arXiv: 2006.16067, 2020.
62. M. Rowe, N. Krishnamurthi and P. Ma, “Patch-based real time road object detection using YOLOv3,” Stanford University, CA, 2018. [Online]. Available: https://cs230.stanford.edu/projects_fall_2018/reports/12418781.pdf.
63. Abe · 55th Int. Congress of Meat Science and Technology · 2009
64. Kim · X-ray apparatus for detecting a flaw of small sized article continuously · 2015