A SVM Based Method to Detect Color Shift Defects in IC Packages

Automated Visual Inspection (AVI) is an essential part in the manufacturing process of Integrated Circuit (IC) packages. Contamination a common defect type found in IC packages appears as a shift in color. One of the main difficulties of this kind of inspection is manual parameter tuning, considering the fact that metallic areas change their colors from product to product, and depending on the IC package material and AVI System lighting. The main target of this paper is to overcome this limitation by automating setting the decision rule, which is very difficult due the large number of parameters and their multidimensional behavior. For this purpose a novel parameter learning system based on Support Vector Machines (SVM) is proposed here to solve this problem.

• Type: Conference
• Year: 2007
• Where: MVA
• Authors: R. M. C. B. Ratnayake, Craig Hicks, M. A. Akbari
• Keywords: Computer Vision, Ic Package, IC Package Material, MVA 2007, Parameter