Abstract
GHOUGH is a computer program for detecting instances of a given shape within an image. It may be used for cueing, counting, or mensuration. GHOUGH can find instances that are displaced, rescaled, rotated, or incomplete relative to the shape template. They are detected by computing a “generalized Hough transform’’ of the image edge elements. Each edge element votes for all those instances of the shape that could contain it; the votes are tallied and the best supported instances are reported as likely matches. GHOUGH was contributed to the DARPA Image Understanding Testbed at SRI by the University of Rochester. This report summarizes applications for which GHOUGH is suited, the history and nature of the algorithm, details of the Testbed implementation, the manner in which GHOUGH is invoked and controlled, the types of results that can be expected, and suggestions for further development. The scientific contributions of this technical note are the analysis of GHOUGH’s parameter settings and performance characteristics.
Share this
