A hybrid system for handprinted word recognition

The authors describe an alternative approach to handprinted word recognition using a hybrid of procedural and connectionist techniques. They utilize two connectionist components, which are to concurrently make recognition and segmentation hypotheses, and to perform refined recognition of segmented characters. Both networks are governed by a procedural controller which incorporates systematic domain knowledge and procedural algorithms to guide recognition. A recognition method is presented whereby an image is processed over time by a spatiotemporal connectionist network. The scheme offers several attractive features including shift-invariance and retention of local spatial relationships along the dimensions being temporalized, a reduction in the number of free parameters, and the ability to process arbitrarily long images. Recognition results on a set of real-world isolated zip code digits are comparable to the best reported to date with a 96.0% recognition rate. A pilot implementation of the complete system, and results on overlapping and touching pairs of zip code digits are reported