Handwritten Character Recognition Based on Relative Position of Local Features Extracted by Self-Organizing Maps
Keiichi Horio, Takeshi Yamakawa · Institutional Repositories DataBase (IRDB) · 2007
ABSTRACT. This paper describes a new pattern recognition method which is based on relative position of local features. We use a self-organizing map to detect a position of features, thus the relative position of them are automatically defined based on arrangement of the competing units. The local features are detected using filter constructed by adaptive subspace self-organizing maps.