Recognizing Shapes Via Random Chord Samplings

A brief introduction to various methods of two dimensional shape recognition is given. A procedure for image-plane rotation and translation invariant object recognition is described. The method is based on sampling random chords from the greylevel image of an object and using the texture information associated with these chords to recognize the object. A voting scheme based on a generalized Hough transform performs the recognition. Schemes for organizing random chord shape descriptions and for matching the textural information associated with the chords are presented in terms of the overall voting scheme. Examples are given which show the effectiveness of this technique. Finally, directions of future research are discussed.
Note: Abstract extracted from PDF file via OCR

Hashim, R, and Worthy Martin. "Recognizing Shapes Via Random Chord Samplings." University of Virginia Dept. of Computer Science Tech Report (1986).