Hybrid systems-a key to intelligent pattern recognition

For the solution of complex problems in pattern recognition and more generally in machine intelligence, involving heterogeneous data sources of both numeric and symbolic information, the fundamental design philosophy is to employ hybrid methodologies rather than attempting to produce the solution using a single paradigm. The authors discuss some of the key issues which need to be addressed in integrating heterogeneous methodologies for intelligent solution of pattern recognition problems and some of the design principles found useful. They also describe some proof-of-concept systems developed at LNK Corporation. These systems include an automatic feature extraction and a target recognition system which combines heterogeneous technologies such as neural network paradigms, uncertainty calculi and expert systems, and heterogeneous data sources including optical and synthetic aperture radar and terrain databases