Recognition-based handwritten Chinese character segmentation using a probabilistic Viterbi algorithm

This paper presents a recognition-based character segmentation method for h andwritten Chinese characters. Possible non-linear segmentation paths are i nitially located using a probabilistic Viterbi algorithm. Candidate segment ation paths are determined by verifying overlapping paths, between-characte r gaps, and adjacent-path distances. A segmentation graph is then construct ed using candidate paths to represent nodes and two nodes with appropriate distances are connected by an are. The cost in each are is a function of ch aracter recognition distances, squareness of characters and internal gaps i n characters. After the shortest path is detected from the segmentation gra ph, the nodes in the path represent optimal segmentation paths. In addition , 125 text-line images are collected from seven form documents. Cumulativel y, these text-lines contain 1132 handwritten Chinese characters. The averag e segmentation rate in our experiments is 95.58%. Moreover, the probabilistic Viterbi algorithm is modified slightly to extra ct text-lines from document pages by obtaining non-linear paths while gaps between text-lines are not obvious. This algorithm can also be modified to segment characters from printed text-line images by adjusting parameters us ed to represent costs of arcs in the segmentation graph. (C) 1999 Elsevier Science B.V. All rights reserved.