Extended Fuzzy Hyperline Segment Neural Network for Handwritten Character Recognition
Dipti Pawar · 2012
This paper deals with simple and effective set of features for character representation. These features are computed within regularly placed windows spanning the character bitmap; consist of a combination of average pixel density and measures of local alignment along some directions. NIST database and Devnagari digit databases are used for experimentation. In NIST database, training set consists of 60,000 patterns and testing set consist of 10,000 patterns. Devnagari digit database consist of 500 patterns. It is divided in two parts. Training set consists of 300 patterns and testing set consists of 200 patterns. These features used in conjunction with Extended Fuzzy Hyperline Segment Neural network (EFHLSNN). The performance of EFHLSNN is found to be superior compared to FHLSNN with respect to training time, recall time per pattern and recognition rate. Index Terms— Fuzzy Neural Network, Feature Extraction, Hyperline Segment, Handwritten Character Recognition