The present book represents a partial record of the 1991 Workshop on "Learning and Geometry" hosted by the University of Maryland and the Center for Night Vision. As the fields of Geometry Theorem Proving and Computational Learning are at present quite mature, being able to tackle real-life problems, it has become an interesting research topic to try to integrate them, especially in order to deal with applications in which no one of these two approaches has been successful on its own. Such applications include interpreting data produced by a variety of sensors (current vision techniques based on computational geometry being able to extract features from data, but failing in the task of recognition of the sensed objects).
edited by David Kueker and Carl Smith