Recognition of courtesy amounts on bank checks based on a segmentation approach

Recognition of Courtesy Amounts on Bank Checks based on a Segmentation Approach Li Qiang Shang X segmentation based courtesy amount recognition (CAR) system is presented in this thesis. This system consists of four modules, which are preprocessing, segmentation, recognition and syntacticai analysis. The noise is fht removed in the preprocessing module. Then, the segmentation module seeks to extract characters (digits, pmctuation marks and so on) fiom the the numeral strings. The recognition module aims at ciassifjing each pattern into one of the ten numerals ('O' '9') and double zeros ('OO'), rejecting ambiguous patterns. Finaily, the syntactical analysis module parses the recognition d t s to provide an acceptable courtesy amount. Two sets of features, shape features and spatial features, are developed to locate the punctuation marks. In addition, a hypotheses-then-verification approach for segmentation of the numeral strings is introduced. The emphasis is on hding ali possible segmentation hypotheses and then verify them. A twdevel segmentation module is proposed, nameiy the globaI segmentation IeveI and the local segmentation IeveI. At the global segmentation level, the intent is to iind the potentiai broken numerals and group them together. On the contrary? the local segmentation level seeks to split the touching digits. Two dassiîiers are c o m b i i into the recognition module. The isolateci digit classifier divides the input patterns into ten numerals ('O' '9'), wMe the holistic doubIe zem classifier intends to recognize the cursive and touching double zeros as an atomic symbol. The proposed courtesy amonnt recognition system has been trained on the database of CENPARi'i' checks and tested on the database of real checks. The system reads 66.5% real checks c o d y at O% misreading rate.