Uncertain Classification Decision for Handwritten Numerals Recognition

碩士 === 淡江大學 === 資訊管理學系 === 87 === A uncertain classification decision system for handwritten numerals recognition is proposed. A thinning algorithm is first used to obtain the skeleton of a numeral. A set of feature points are then detected. The skeleton can be decomposed into geometric primitives....

Full description

Bibliographic Details
Main Authors: Tai-Yuan Hsiao, 蕭泰源
Other Authors: Hung-chang Lee
Format: Others
Language:zh-TW
Published: 1999
Online Access:http://ndltd.ncl.edu.tw/handle/30689272295921333832
Description
Summary:碩士 === 淡江大學 === 資訊管理學系 === 87 === A uncertain classification decision system for handwritten numerals recognition is proposed. A thinning algorithm is first used to obtain the skeleton of a numeral. A set of feature points are then detected. The skeleton can be decomposed into geometric primitives. Finally, we use the relationships between the primitives to recognize the numeral. For the digits which have the similar primitives, a membership function is used to decide the probability of each digit. The handwritten numerals extracted from the NIST Special Database 19 are used to test the system. The correct rate is 88.72%.