This paper presents the candidate elimination implementation of the version space strategy for classification of photographic data. It is shown that very accurate classification is easily achieved and that only a small number of training samples are needed to generate the rules.
Kaminsky, Edit J., and Willie G. Brown, "Image Classification using Version Spaces," in lASTED Signal and Image Processing Conf. Proc. (SIP'95), (N. M. Namazi, editor), Las Vegas, NV, Nov. 20-23,1995, pp. 284-287.