EUVIMEDEUVIMED entdecken
Crossref

107‐3: Unsupervised OLED Mura Defect Detection via Denoising Diffusion Probabilistic Models

Wei-Ting Tsao, Shang-Yu Su, Feng-Ting Pai, Chia-Hsin Tung, Wen-Pin Tsai

SID Symposium Digest of Technical Papers · 2026-06 · 57 (1) · 1491-1494

DOI: 10.1002/sdtp.70001_403 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SID Symposium Digest of Technical Papers, online erschienen am 11.09.2026. Verfasst von Wei-Ting Tsao und 4 weiteren Autor:innen. Die Arbeit nennt 8 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In this paper, we address the problem of OLED mura detection and propose an unsupervised generative AI approach to achieve robust defect identification. Our method leverages a generative model to localize the positions of mura defects without the need for labeled data Furthermore, we introduce a discriminative quantitative metric that enables objective assessment of defect severity. The proposed framework demonstrates both precise localization and effective quantification, offering a comprehensive solution for OLED display quality inspection.

Zitieren

APA

Tsao, W., Su, S., Pai, F., Tung, C., & Tsai, W. (2026). 107‐3: Unsupervised OLED Mura Defect Detection via Denoising Diffusion Probabilistic Models. SID Symposium Digest of Technical Papers, 57(1), 1491-1494. https://doi.org/10.1002/sdtp.70001_403

Vancouver

Tsao W, Su S, Pai F, Tung C, Tsai W. 107‐3: Unsupervised OLED Mura Defect Detection via Denoising Diffusion Probabilistic Models. SID Symposium Digest of Technical Papers. 2026;57(1):1491-1494. doi:10.1002/sdtp.70001_403

BibTeX
@article{Tsao2026,
  author = {Tsao, Wei-Ting and Su, Shang-Yu and Pai, Feng-Ting and Tung, Chia-Hsin and Tsai, Wen-Pin},
  title = {107‐3: Unsupervised OLED Mura Defect Detection via Denoising Diffusion Probabilistic Models},
  journal = {SID Symposium Digest of Technical Papers},
  year = {2026},
  volume = {57},
  number = {1},
  pages = {1491--1494},
  doi = {10.1002/sdtp.70001_403}
}
Kostenlos starten →