REASONING ABOUT ICONIC DATA IN ARTIFICIAL VISION
This paper proposes a 3-level architecture for an artifical vision system, integrating low level processing and high level reasoning features. The interface between different levels is realized by means of an iconic data base, where a 3-D description of the scene is stored and iteratively refined. The three levels considered are: early processing (numerical algorithms for peripheral elaboration), perceptual reasoning (grouping the data obtained from the previous level on the basis of some perceptual principle), conceptual reasoning (knowledge- driven processing: recognition, objects, functionality etc.). The implementation of some working modules is presented along with the iconic database and some examples discussed.
Bibliographic Reference: 2ND INTERNATIONAL SYMPOSIUM ON OPTICAL & ELECTRO-OPTICAL SCIENCE ENGINEERING & INSTRUMENT DISPLAY, CANNES (FRANCE), DEC. 2-6, 1985 WRITE TO CEC LUXEMBOURG, DG XIII/A2, POB 1907 MENTIONING PAPER E 32475 ORA
Availability: Can be ordered online
Record Number: 1989124069200 / Last updated on: 1987-01-01
Available languages: en