Metraipolis › archive › dc_414ef29bdcc8
Beyond CLIP: Calibrated (Few-Shot) Defect Detection With System One Decision Models — arXiv:2610.41199
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.41199
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-02T07:01:02.190Z
- SHA-256 of the archived text
82d30da51a4cf842fd924a84a4a972024e88efc10d689da9a238e43dec449993- Stable URL of this page
https://metraipolis.com/documento/dc_414ef29bdcc8
How to cite
Beyond CLIP: Calibrated (Few-Shot) Defect Detection With System One Decision Models — arXiv:2610.41199 — arXiv (en), https://arxiv.org/abs/2610.41199 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_414ef29bdcc8 (Metraipolis archive dc_414ef29bdcc8, sha256 82d30da51a4cf842fd924a84a4a972024e88efc10d689da9a238e43dec449993)
BibTeX
@misc{dc414ef29bdcc8,
title = {Beyond CLIP: Calibrated (Few-Shot) Defect Detection With System One Decision Models — arXiv:2610.41199},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.41199},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_414ef29bdcc8; sha256 82d30da51a4cf842fd924a84a4a972024e88efc10d689da9a238e43dec449993},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41199. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Cameras and vision sensors are widely used for product and process inspection in manufacturing environments. Traditionally, this process required thousands of images of both nominal and defective products to train such models. Recent research shows that vision transformer (ViT) models, such as CLIP, can reduce this requirement to 50-100 images per nominal and defective class. However, collecting even 50 defective images can be difficult in high-yield or low-volume manufacturing processes. In this paper, we examine the utility of two recently proposed structured decision models (Jev and Clef-Flash) in applications using as few as k=2 reference images per class at inference. In the process, we provide three main contributions: (a) a framework for k-shot learning of these models for image classification/inspection applications; (b) an agentic approach for automatically refining the text descriptions that guide image classification in these decision models; and (c) an approach to calibrate the resulting defect probabilities to improve their reliability for inspection decisions. We test these contributions in five manufacturing case studies. Using two reference images per class, both Jev and Clef-Flash outperform CLIP's test accuracies with the same references (0.75-1.00 [Jev] and 0.71-0.99 [Clef-Flash] vs 0.67-0.84 [CLIP]). Furthermore, we show that our agentic refinement of text descriptions significantly improves accuracy across 6 of 8 model-application combinations, and our calibration approach reduces the Brier score in 13 of the 15 model-application configurations. To support the reproducibility of our work, we provide our replication code and have created the R "decisions" package for evaluating additional models. We also provide a publicly available web app that allows practitioners to inspect their own images with no code. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.41199 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-02T06:56:58.379Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).