A pre-alarm system for visual fatigue in an IC packaging factory

Before shipping the product to customer, maintenance is a back-end department of integrated circuit (IC) production, and yield rate measures whether the product batch is accepted or not. A correlation analysis showed a positive relationship between yield rate and visual fatigue, so reducing visual fatigue can improve operators' comfort as well as the yield rate. In this study, Group Method and Data Handling (GMDH) of neural network was applied to build a visual fatigue prediction model which was implemented in the maintenance section of an IC packaging factory. The implementation result showed that visual fatigue was reduced, and yield rate also improved simultaneously.