Off-line system for recognition of free written zipcodes

Off-line system for the recognition of free written zip-codes is presented in the paper. The system is built upon five main components. A division module separates the input zip-codes into isolated numeral. The preprocessing module do all the normalizing, noise removing and thinning. The feature division gives several kinds of feature description and extraction. We have two stages of classification, i.e., coarse and detailed. The former uses the inference procedure based on local structure, with Fuzzy pre-classifier being set at the front of the recognition system. The later uses distance matching on subset of prototypes, discrete wavelet transform on some local curve segment, single-layered BP network on whole binary image, which reduces significantly the number of prototypes to be matched with the input digit sample. Experimental results have showed the system is efficient on zip-codes collected from practical envelopes.