Self inspection of integrated circuits pattern defects using support vector machines

摘要

At various steps of integrated circuit manufacturing, defects need to be inspected to ensure process integrity. It is desirable to detect the defects based on the captured optical image itself without comparing to any reference image. Such a methodology of defect inspection without any reference image is called defect self inspection. In this article, we investigate the defect self inspection using the support vector machines, a pattern classification methodology, and present detection sensitivity of this approach.