Feature Selection for Unsupervised Anomaly Detection and Localization Using Synthetic Defects

Lars Heckler, Lars Heckler, Rebecca König

2024

Abstract

Expressive features are crucial for unsupervised visual Anomaly Detection and Localization. State-of-the-art methods like PatchCore or SimpleNet heavily exploit such features from pretrained extractor networks and model their distribution or utilize them for training further parts of the model. However, the layers commonly used for feature extraction might not represent the optimal choice for reaching maximum performance. Thus, we present the first application-specific feature selection strategy for the task of unsupervised Anomaly Detection and Localization that identifies the most suitable layer of a pretrained feature extractor based on the performance on a synthetic validation set. The proposed selection strategy is applicable to any feature extraction-based AD method and may serve as a competitive baseline for future work by not only outperforming single-layer baselines but also features ensembled from multiple layer outputs.

Paper Citation

in Harvard Style

Heckler L. and König R. (2024). Feature Selection for Unsupervised Anomaly Detection and Localization Using Synthetic Defects. In Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3: VISAPP; ISBN 978-989-758-679-8, SciTePress, pages 154-165. DOI: 10.5220/0012385500003660

in Bibtex Style

@conference{visapp24,
author={Lars Heckler and Rebecca König},
title={Feature Selection for Unsupervised Anomaly Detection and Localization Using Synthetic Defects},
booktitle={Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3: VISAPP},
year={2024},
pages={154-165},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012385500003660},
isbn={978-989-758-679-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3: VISAPP
TI - Feature Selection for Unsupervised Anomaly Detection and Localization Using Synthetic Defects
SN - 978-989-758-679-8
AU - Heckler L.
AU - König R.
PY - 2024
SP - 154
EP - 165
DO - 10.5220/0012385500003660
PB - SciTePress