Integrating content-based image retrieval and deep learning to improve wafer bin map defect patterns classification

ABSTRACT

Defect dies scattering on semiconductor wafer bin maps (WBM) tends to form specific patterns that point to particular manufacturing problems. The distribution of defect patterns from the shop floor is often highly imbalanced, leading to the challenge of having insufficient data about defect pattern types when building deep learning classification models. The method for completing such analysis in a timely manner with limited data is of critical interest. This study developed a method for applying content-based image retrieval (CBIR) and convolutional neural networking (CNN) to WBM defect patterns classification to solve the data imbalance problem and to improve accuracy when using relatively a small quantity of data. In this research, 3,600 WBMs featuring 12 defect pattern types were selected from the WM-811 K dataset for empirical validation. Using only 1,400 CNN training data elements, the overall classification accuracy reached 98.44%.

Graphical abstract

KEYWORDS:

• Semiconductor
• Wafer bin map
• defect pattern
• Content-Based Image Retrieval (CBIR)
• pattern recognition
• Convolutional Neural Network (CNN)

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Ministry of Science and Technology, Taiwan [MOST 109-2628-E-007-002-MY3].

Notes on contributors

Ming-Chuan Chiu

Dr. Ming-Chuan Chiu is a Professor, in the Department of Industrial Engineering & Engineering Management, National Tsing Hua University (NTHU), Taiwan. He obtained his BS and MS degrees Industrial Engineering & Engineering Management from National Tsing Hua University (NTHU) in 1997 and 1999, respectively. Before joining NTHU, he served at The School of Engineering Design, Technology, and Professional Programs (SEDTAPP) at The Pennsylvania State University as instructor for one year and worked in industry as chief engineer for six years. He received his PhD degree in the Department of Industrial and Manufacturing Engineering at The Pennsylvania State University in 2010. Dr. Chiu joined the faculty of Dept. Industrial Engineering and Engineering Management, National Tsing Hua University (NTHU) in August 2011. He has served as an associate professor since August 2016. He is also the founder of Innovation Engineering Laboratory (IELAB) at National Tsing-Hua University. Dr. Chiu was Sayling Wen's Award for Outstanding Young Researcher in Service Science in 2013 and receipt of MOST Outstanding Young Scholar Grant in 2014~2017. He serves as associate editor of International Journal of Industrial Engineering: Theory, Applications and Practice (SCI) since 2016 and editorial board of Advanced Engineering Informatics (SCI) since 2019. Dr. Chiu has been acting as the PI and co-PI for more than 20 projects from the government (including Ministry of Science and Technology (MOST), Ministry of Education (MOE), Industrial Technology Research Institute (ITRI), Institute for Information industry (III)) and industries. His research interests focus on Sustainable Design, Service Innovation, Product Service System Design, and Smart Manufacturing. The aim of the above interests is to solve problems in product, service, and system development stage using systems thinking.

Yen-Han Lee

Yen-Han Lee received her MS degree in Innovation Engineering Laboratory (IELAB) in the Department of Industrial Engineering and Engineering Management at National Tsing-Hua University in winter of 2019. Her research interests focus on development of Wafer Map Defect Patterns Classification. The aim of the above interests is to solve problems in semiconductor industry to improve competitive advantage.

Tao-Ming Chen

Tao-Ming Chen received his MS degree in Innovation Engineering Laboratory (IELAB) in the Department of Industrial Engineering and Engineering Management at National Tsing-Hua University in summer of 2021. His research interests focus on development of Mixed-Type Wafer Maps Defect Patterns Classification. The aim of the above interests is to solve problems in semiconductor industry to improve competitive advantage.