CONNECTED CHARACTER-RECOGNITION IN CURSIVE HANDWRITING USING SELECTIVE ATTENTION MODEL WITH BEND PROCESSING

As one of the systems to recognize the cursive handwritten English connected character sequence, Imagawa et al. proposed a system with the selective attention mechanism. The system of Imagawa et al., however, does not have a very high recognition ability. It is a relatively small-scale system, where the character category to be recognized is composed of 5 characters. This study extends the recognition system proposed by Imagawa et, al., and a system with a higher recognition ability is constructed. One of the authors notes that a part of the ''model for selective attention'' is similar to the pattern recognition system ''neocognitron.'' With respect to the neocognitron, it is reported that the recognition ability is improved by introducing the bend detection circuit. Consequently, the authors constructed a system, where the bend processing circuit is introduced into the system by Imagawa et al. Various test patterns are given to the constructed system, and it is verified by executing a computer simulation that the proposed system is effective in the recognition of the handwritten connected character sequence.