Redistribution Layer Defect Classification Using Computer Vision Techniques And Machine Learning

Abstract

In the semiconductor industry, defects are yield killers and the detection/classification of which can be expensive as well as time consuming. To overcome this challenge, we propose a solution involving Computer Vision Techniques and Machine Learning to accomplish defect binning procedure in typical wafer-level packaging scenario, focusing on 2um L/S redistribution layer (RDL) features. With this approach, inspection cycle time is reduced, thereby driving faster product development.