Visual Probe Mark Inspection, Using Hardware Implementation of Artificial Neural Networks, in VLSI Production

摘要: As a result of their adaptability, artificial neural networks present good solutions for permanently increasing range industrials problems. So, if usefulness has already been confirmed, very few papers deal with real applications this kind technology. Our goal is to based solution that we have developed visual inspection in VLSI production the IBM Essonnes plant. The main characteristics such systems are real-time control and high reliability detection classification tasks. presented system on ZISC©, an hardware implementation Restricted Coulomb Energy algorithm K-Nearest Neighbor algorithm. application inspect vias probe damage during wafer tests: each via analyzed classified (good impact, bad impact or absence impact). First results really encouraging show efficiency manufacturing environment.