Towards Total Recall in Industrial Anomaly Detection

By Karsten Roth, Latha Pemula, Joaquin Zepeda, Bernhard Schölkopf, Thomas Brox, Peter Gehler. Computer Vision and Pattern Recognition, 2021

1,710 citations recorded by the source corpus.

Being able to spot defective parts is a critical component in large-scale industrial manufacturing. A particular challenge that we address in this work is the cold-start problem: fit a model using nominal (non-defective) example images only. While handcrafted solutions per class are possible, the goal is to build systems that work well simultaneously on many different tasks automatically. The best peforming approaches combine embeddings from ImageNet models with an outlier detection model. In this paper, we extend on this line of work and propose PatchCore, which uses a maximally representative memory bank of nominal patch-features. PatchCore offers competitive inference times while achieving state-of-the-art performance for both detection and localization. On the challenging, widely used MVTec AD benchmark PatchCore achieves an image-level anomaly detection AUROC score of up to 99.6%, more than halving the error compared to the next best competitor. We further report competitive results on two additional datasets and also find competitive results in the few samples regime. Code: github.com/amazon-research/patchcore-inspection.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/23ad8fc48530ce366f8192dfb48d0f7df1dba277

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Ziwei Liu
2. Yixuan Li
3. Kaiyang Zhou
4. Jingkang Yang
5. Bo Li
6. Wayne Zhang
7. Haoqi Wang
8. Yiyou Sun

Citations in the corpus

• FastFlow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flows (2021, methodology): “2021), Patch Core (Roth et al. 2021), CFlow (Gudovskiy, Ishizaka, and Kozuka 2021) under the metrics of image-level AUC and pixel-level AUC.”
• Deep Industrial Image Anomaly Detection: A Survey (2023, neutral): “PatchCore [68], SPADE [63], PaDim [73] conduct the ablation study on 16 normal training samples.”
• EfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latencies (2023, methodology): “[50] Karsten Roth, Latha Pemula, Joaquin Zepeda, Bernhard Schölkopf, Thomas Brox, and Peter Gehler.”
• AnomalyGPT: Detecting Industrial Anomalies using Large Vision-Language Models (2023, methodology): “The design of the decoder is primarily inspired by PatchCore [23], WinCLIP [11], and APRIL-GAN [2].”
• CFA: Coupled-Hypersphere-Based Feature Adaptation for Target-Oriented Anomaly Localization (2022, background): “Recently, [4, 5, 16] designed memory banks using pre-trained CNNs with large datasets such as ImageNet [6] and achieved state-of-the-art (SOTA) performance.”
• DeSTSeg: Segmentation Guided Denoising Student-Teacher for Anomaly Detection (2022, methodology): “(b) Downsample with bilinear interpolation, then floor all values between (0, 1) to zero [13, 25].”
• OpenOOD: Benchmarking Generalized Out-of-Distribution Detection (2022, background): “PatchCore [57] samples ID features into a memory bank and performs the nearest neighbor searching to discover anomalies.”
• SimpleNet: A Simple Network for Image Anomaly Detection and Localization (2023, methodology): “Feature Extractor acquires local feature as in [22].”
• AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly Detection (2023, neutral): “Comparison with SOTA full-shot methods In this section, we are interested in the performance gap between AnomalyCLIP and the recently published SOTA full-shot methods, such as Patch-Core (Roth et al., 2022) and RD4AD (Deng & Li, 2022).”
• Characterizing Generalization under Out-Of-Distribution Shifts in Deep Metric Learning (2021, background): “However, in practice, achieving effective OOD generalization is more challenging than in-distribution [28, 12, 21, 49, 31, 55].”
• SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentation (2022, methodology): “For both setups, existing state-of-the-art methods for surface anomaly detection commonly leverage supervised representations pre-trained on ImageNet [16], either as feature extractors [14,33] or as initialization for fine-tuning on the target dataset [26,44].”
• Revisiting Reverse Distillation for Anomaly Detection (2023, methodology): “FastFlow [39] 80.40 CFA [20] 98.25 PatchCore [27] 99.70 RD [8] 99.36 RD++ 99.73 Figure 5.”