ANALYSIS SYSTEM OF MIXED DOCUMENTS CONSISTING OF HANDWRITTEN KOREAN/ALPHANUMERIC TEXTS AND GRAPHIC IMAGES
Chang Bum Lee, Donggyu Sim, In Kwon Kim, Young Kug Ham, Rae‐Hong Park · 1994
bound algorithm based on DP matching costs. Also we recognize alphanumeric characters using scvcral robust This paper proposes an effective recognition system features. Finally, to correct wrong rccognition results, we which recognizes the mixed document consisting of use a dictionary and knowledge employed in a recognition handwritten Knreadalphanumeric texts and graphic step. images. In the preprocessing step, an input image is binarized by the proposed thresholding scheme, then PREPROCESSING graphic and character regions are separated by using The proposed algorithm consists of the preproccssing chain codes of connected components. In the character and recognition steps. In the preprocessing step, each