A complete and scalable architecture for 3D model-based vision

Abstract

As a follow-up to the work of R. M. Bolle et al. (Proc. DARPA Image Understanding Workshop, May 1989, pp. 471-478), an overview is given of subsequent work. The focus of the ongoing work is the development of an experimental vision system for recognition of 3-D objects. The thrust of this vision system is to investigate techniques that may lead to a system that scales with the size of the problem; by the size of the problem, the authors mean the complexity of the scene -- the number of objects in the scene and the number of objects in the database, i.e., the number of objects that the system can recognize. One important aspect of the work is representing 3-D shapes and objects, and storing these representations in object-model databases. Another aspect is the control of the search, that is, finding good intermediate representations of the data and using these representations to index into the object-model database.