Feature-based defect detection

Apparatus for inspecting a patterned substrate that comprises:

• a computer system;
• and
• instructions for controlling the computer system:
• to prepare a reference image and a test image of at least a portion of the patterned substrate;
• to extract features from the reference image and extract features from the test image;
• to match features of the reference image and features of the test image;
• and
• to compare features of the reference image and of the test image to identify defects;
• wherein comparing features of the reference image and of the test image to identify defects comprises:
• computing feature properties in the reference image;
• computing feature properties in the test image;
• and comparing the computed feature properties in the test image with the computed feature properties in the reference image;
• wherein the computed feature properties comprise:
• area, size, center of gravity, average intensity, sum of intensity, and standard deviation of intensity;
• wherein comparing the computed feature properties comprises:
• calculating feature properties comparison statistics that include one or more of differences, ratios, greater-than or less-than functions, means, standard deviations, linear regressions, and adaptive thresholds;
• and
• determining whether the feature properties comparison statistics meet predetermined defect criteria.