Extraction of Knowledge from Data: An Evaluation of Intelligent Knowledge Based Systems (IKBS)

摘要

Automatic target recognition (ATR) by a sensor-computer system, and analogies with human visual perception are discussed. The best approach to ATR modeling may not involve detailed replication of the brain's processing steps in recording the sensory input. However, the general logic of the brain is relevant, including ways of eliminating redundant information, ways of using selected feature extractors as building blocks at each level of restriction such as position on the generality of the feature, and the use of fuzzy logic to provide continuously variable levels of stringency in the encoding process. The early stages of visual processing were incorporated in the VISIVE program, enabling information on important details in a scene to be extracted.