Character Recognition in Scene Images by Fast Deformable Template Matching
Masakazu Iwamura, Shinichiro Omachi, Hirotomo Aso · 2006
To recognize characters in scene images, a segmentation method is generally required for pre-processing. However, there are problems that traditional methods are not able to segment charac- ter which does not constitute a string and character which is connected to other characters. Moreover, it is known that images in scene images suffer from perspective distortion. In this paper, we propose a new fast segmentation method of characters including the isolated and the touching ones with robustness for perspective distortion and deformations. By applying the proposed method to scene images, it is confirmed that the method is robust against perspective distortion and deformations, and the method