An urn theoretical approach towards modelling of thinning algorithms

The urn theoretic model introduced has been applied in deducing some interesting results for image processing. One of the most intuitively appealing applications of urn models is in the area of stochastic learning theory. Using this concept, the performance of a class of pattern thinning algorithms in image processing has been analysed. The performance can be visualized as the probability of a correct response in a multiple choice problem. A simple replacement model has been used for the performance analysis of some thinning algorithms.<>