Intelligent defect diagnosis method

An intelligent defect diagnosis method for manufacturing fab, comprising:

• receiving pluralities of defect data from at least one defect inspection tool, pluralities of design layouts generated from a profile design system, and pluralities of fabrication data produced in the fab;
• analyzing the defect data, design layouts, and the fabrication data by a defect analysis system, wherein the analyzing step further contains the following sub-steps:
• segmenting and grouping the design layouts into pluralities of multi-pattern group cells to construct Layout Pattern Group (LPG) cell based pattern groups;
• introducing the defect data;
• segmenting defect image into pluralities of defect and pattern contours;
• mapping the defect data to each multi-pattern group cells to form the LPG based defect composite pattern groups, and identifying the frequent failure defect layout pattern;
• performing coordinate conversion and image pattern match between image pattern contour and design layout for coordinate correction;
• fulfilling Critical Area Analysis (CAA) with defect contour, pattern contour or design layout polygon, and obtaining corresponding defect yield;
• classifying the defect type of defect data through defect image classification analysis.