# PANDA: Adapting Pretrained Features for Anomaly Detection and Segmentation

arXiv: https://arxiv.org/abs/2010.05903 · HTML: https://mlsanity.com/paper/2010.05903/

## Key facts

- Cited by 175 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Oct 2020.

## Built on by (most cited first, 30 of 175)

- [Generalized Out-of-Distribution Detection: A Survey](https://mlsanity.com/paper/2110.11334/) (2021)
- [Anomaly Detection via Reverse Distillation from One-Class Embedding](https://mlsanity.com/paper/2201.10703/) (2022)
- [A Unified Model for Multi-class Anomaly Detection](https://mlsanity.com/paper/2206.03687/) (2022)
- [ADBench: Anomaly Detection Benchmark](https://mlsanity.com/paper/2206.09426/) (2022)
- [FastFlow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flows](https://mlsanity.com/paper/2111.07677/) (2021)
- [Deep Industrial Image Anomaly Detection: A Survey](https://mlsanity.com/paper/2301.11514/) (2023)
- [Catching Both Gray and Black Swans: Open-set Supervised Anomaly Detection](https://mlsanity.com/paper/2203.14506/) (2022)
- [Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection](https://mlsanity.com/paper/2203.05550/) (2022)
- [Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection](https://mlsanity.com/paper/2405.14325/) (2024)
- [Mean-Shifted Contrastive Loss for Anomaly Detection](https://mlsanity.com/paper/2106.03844/) (2021)
- [A Unified Anomaly Synthesis Strategy with Gradient Ascent for Industrial Anomaly Detection and Localization](https://mlsanity.com/paper/2407.09359/) (2024)
- [Explicit Boundary Guided Semi-Push-Pull Contrastive Learning for Supervised Anomaly Detection](https://mlsanity.com/paper/2207.01463/) (2022)
- [Explainable Deep Few-shot Anomaly Detection with Deviation Networks](https://mlsanity.com/paper/2108.00462/) (2021)
- [Deep One-Class Classification via Interpolated Gaussian Descriptor](https://mlsanity.com/paper/2101.10043/) (2021)
- [Towards Scalable 3D Anomaly Detection and Localization: A Benchmark via 3D Anomaly Synthesis and A Self-Supervised Learning Network](https://mlsanity.com/paper/2311.14897/) (2023)
- [Anomaly Detection in 3D Point Clouds using Deep Geometric Descriptors](https://mlsanity.com/paper/2202.11660/) (2022)
- [Hierarchical Vector Quantized Transformer for Multi-class Unsupervised Anomaly Detection](https://mlsanity.com/paper/2310.14228/) (2023)
- [IM-IAD: Industrial Image Anomaly Detection Benchmark in Manufacturing](https://mlsanity.com/paper/2301.13359/) (2023)
- [ReContrast: Domain-Specific Anomaly Detection via Contrastive Reconstruction](https://mlsanity.com/paper/2306.02602/) (2023)
- [Collaborative Discrepancy Optimization for Reliable Image Anomaly Localization](https://mlsanity.com/paper/2302.08769/) (2023)
- [SoftPatch: Unsupervised Anomaly Detection with Noisy Data](https://mlsanity.com/paper/2403.14233/) (2024)
- [Latent Outlier Exposure for Anomaly Detection with Contaminated Data](https://mlsanity.com/paper/2202.08088/) (2022)
- [PAD: A Dataset and Benchmark for Pose-agnostic Anomaly Detection](https://mlsanity.com/paper/2310.07716/) (2023)
- [Multimodal Industrial Anomaly Detection by Crossmodal Feature Mapping](https://mlsanity.com/paper/2312.04521/) (2023)
- [Improving neural network representations using human similarity judgments](https://mlsanity.com/paper/2306.04507/) (2023)
- [Exposing Outlier Exposure: What Can Be Learned From Few, One, and Zero Outlier Images](https://mlsanity.com/paper/2205.11474/) (2022)
- [One-shot Unsupervised Domain Adaptation with Personalized Diffusion Models](https://mlsanity.com/paper/2303.18080/) (2023)
- [SQUID: Deep Feature In-Painting for Unsupervised Anomaly Detection](https://mlsanity.com/paper/2111.13495/) (2021)
- [Active Learning Helps Pretrained Models Learn the Intended Task](https://mlsanity.com/paper/2204.08491/) (2022)
- [Focus Your Distribution: Coarse-to-Fine Non-Contrastive Learning for Anomaly Detection and Localization](https://mlsanity.com/paper/2110.04538/) (2021)

## Abstract (arXiv)

Anomaly detection methods require high-quality features. In recent years, the anomaly detection community has attempted to obtain better features using advances in deep self-supervised feature learning. Surprisingly, a very promising direction, using pretrained deep features, has been mostly overlooked. In this paper, we first empirically establish the perhaps expected, but unreported result, that combining pretrained features with simple anomaly detection and segmentation methods convincingly outperforms, much more complex, state-of-the-art methods. In order to obtain further performance gains in anomaly detection, we adapt pretrained features to the target distribution. Although transfer learning methods are well established in multi-class classification problems, the one-class classification (OCC) setting is not as well explored. It turns out that naive adaptation methods, which typically work well in supervised learning, often result in catastrophic collapse (feature deterioration) and reduce performance in OCC settings. A popular OCC method, DeepSVDD, advocates using specialized architectures, but this limits the adaptation performance gain. We propose two methods for combating collapse: i) a variant of early stopping that dynamically learns the stopping iteration ii) elastic regularization inspired by continual learning. Our method, PANDA, outperforms the state-of-the-art in the OCC, outlier exposure and anomaly segmentation settings by large margins.