FAST IMAGE SEGMENTATION FOR SOME MACHINE VISION APPLICATIONS.

Abstract

The use of an image contrast measure for producing binary segmentations of images in a certain class of applications is described. This method is well-suited for fast pipeline implementations, because the contrast measure uses only two local features in the image. To eliminate segmentation noise, segmentations have been postprocessed using binary morphological operations. This method has been applied to three different microelectronics inspection problems, with consistently good results, and experimental results from each of these applications are presented. Also, this technique is discussed in terms of the theory of polynomial classifiers.