A training method of dictionary characters is presented forhandwritten character recognition by the model based on relaxationapproach. The dictionary characters are constructed by analyzing thestructure of several training characters using curved line approximationand relaxation matching. The performance of the constructed dictionarycharacters was verified by the experiment. 956 (chars.) × 20(sets) in the standard database ETL-8 (Japanese KANJI and HIRAGANA) areused as sample characters for training and 10 (sets) are used as testcharacters (5 known sets and 5 unknown sets). The mean of the number ofconstructed dictionary characters is 3.81 for each category, andrecognition rate is 99.96% on the known sets and 99.73% on the unknownsets
展开▼
