Inductive Learning of Primitive Shape Features of Closed Contours

A method for inductivelearningof primitive features is proposed. Primitive features are well investigated and they are extracted in bottom-up manner fiom training set of patterns by a segmentation method based on genetic algorithm. Merging is applied to selected primitive candidates to filter out their redundancy as similarity and inclusion. Experimental results with map figures are shown.

• Type: Conference
• Year: 1996
• Where: MVA
• Authors: Ichiro Murase, Shun'ichi Kaneko, Satoru Igarashi
• Keywords: Bottom-up Manner Fiom, Inductivelearningof Primitive Features, MVA 1996, MVA 2007, Primitive Features