Automatic tax-sheet input system using handwritten numeral recognition
Pyeoung Kee Kim, Hankyu Lim, Kyu Young Rho, Kang Yoon Lee, Seong Mook Kim · 2002
This paper presents an automatic tax-sheet input system (ATSIS) using a structural numeral recognition. Using an ATSIS is much needed because it helps reduce data input cost and increases the correctness of the input data. A numeral is represented as sequence of stroke codes and feature points. We classify confusing numerals using separate fuzzy functions. An experiment on sample tax-sheets shows favorable results.