An Efficient Character Segmentation Algorithm for Printed Chinese Documents

The character segmentation technology for printed documents is applied in many fields. This paper proposes an efficient character segmentation algorithm for Chinese printed documents, which is suitable for paper watermarking system. This algorithm is composed of three main steps: connected regions recognition, connected regions merging, and fine-gained segmentation, through what the algorithm succeeds in achieving Chinese character segmentation with high accuracy and high consistent segmentation between the digital version and print-scanned version of images from the same documents. Experiments show the effectiveness of the proposed algorithm.