Syntactic and Structural Pattern Recognition

This is the proceedings of an Advanced Research Workshop within the NATO Scientific Programme held in Sitges, Spain, October 23-25, 1986. Syntactic pattern recognition has a long history of active research and has been used in industrial applications. It was realized early that the methodology of formal languages may not have been appropriate for all problems and one could investigate techniques that emphasize feature analysis and description of the relations using other mathematical tools, such as graph matching. This area of investigation came to be known as structural pattern recognition. Another challenging new direction is offered by the combination of Artificial Intelligence with syntactic and structural pattern recognition (SSPR). The workshop covered not only topics in the mainstream of SSPR but also in the main adjacent areas. The papers in this volume cover matching and parsing techniques mainly in error correcting graphs, grammatical inference under constraints that simplify structural combinations, and logic-syntactic integration. The papers also show significant applications work done in speech recognition (using hybrid techniques), recognition of drawings, feature identification, cryptosystems, and histo-pathology. Moreover, there are several papers on image understanding where SSPR is one of the main techniques applied. In addition to the regular sessions there were Working Groups on 2D and 3D Image Understanding; Speech and Waveform Recognition; Hybrid Methodologies; and Models and Inference. The summaries of the group discussions are incorporated in the volume. One of the interesting aspects treated in the working groups was the relation between representation and techniques, which will probably be the heart of the new generation of pattern recognition techniques.