Abstract: A method of collating and using captured semiconductor-wafer image data in an automated defect analysis. The method includes the steps of receiving image data and, if necessary, converting it to a digital format. Once the data is in pixel-by-pixel form, each pixel is assigned a slope value derived from the direction of the structure edge, if any, on which it lies. The pixel-slope data is then evaluated to determine whether a photo-resist anomaly is present. The method may also include evaluated an average pixel slope value for each inspected wafer. Dependant claims further define the invention to claim an inspection system for employing the method.

Abstract: An apparatus and method for automatic knowledge-based object or anomaly classification is provided by capturing a pixel map of an image and from that generating high level descriptors of the object or anomaly such as size, shape, color and sharpness. These descriptors are compared with sets of descriptors in a knowledge-base to classify the object or anomaly.