Abstract
Wafer bin maps (WBMs) represent the wafer test results in semiconductor processes. In WBMs, identifying the defect patterns are significant because the defect patterns can be used to detect the root causes of a defect. However, WBM researches in a real application suffer from insufficient training data for pattern analysis, so it is hard to apply machine learning techniques. Moreover, it is significant to classify the die, which is the minimum unit in a wafer, for identifying accurate causes, but the previous studies were conducted for the wafers. In this study, we use a variational autoencoder to generate new WBMs samples to build an accurate classifier for defect patterns. We then propose a die-wise defect segmentation system. This system aims to enable classification the cause of defects by using SegNet, a deep convolutional encoder-decoder architecture. To demonstrate the robustness and effectiveness of the proposed methodology, we conducted an experimental evaluation to compare the accuracy of the proposed approach than the ones of two variants of autoencoders. The experimental study confirmed that the proposed approach shows superior performance in the results.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: Journal of the Korean Institute of Industrial Engineers
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
