DOI: 10.1002/sdtp.16485 ISSN:
6‐2: Potential Failure Detection using Unsupervised Clustering and Anomaly Detection
Misuk Kim, Soyeong Park, Seokhyun Yoon, Jung-suk Bae, Nayeon Hwang, Myeonghwa Kim, Jeehun Seo, Dooyoul Lee, Jung-Tae Kang, Jeong-il Yoo- General Medicine
Automated Visual Inspection(AVI) is an effective way for mobile display product to detect defects which may cause quality problem. The rules of determining defects are based on knowledge for an accumulated engineer's experience. However, it is difficult to detect potential failure that may cause in the production process, because we determine failure of panel based on boundary samples. In this study, we propose an analysis method to find abnormalities by quantifying the potential failure with AVI data using unsupervised learning.