Methods for Korean business forms processing

This paper introduces the methods for business forms proessing in Korea. Recognizing the forms of credit card sales slips, bank deposit slips, and insurance application forms will be introduced. This paper is divided into two parts. The first part in section 2 focuses on the forms processing by discussing the issues of form identification, recognition field extraction, and form line removal with character restoration capability. The second part in section 3 focuses on recognizing continuous handwritten numerals extracted from the target application forms. The issues of segmenting touching numerals, classifying each segmented numeral, and improving overall reliability by rejecting garbage patterns will be discussed. Current experimental results are followed in each method introduced.