Handwritten Text Line Segmentation Based on Structural Features

摘要

In this paper, we propose a handwritten text line segmentation of trilingual scripts. Structural features of the script are used for the segmentation of text lines. In the first stage of the proposed method, the connected component and bounding box approach are used as preprocessing. Second stage of the proposed method uses a thinning approach for effective segmentation of the text line. The experiment is conducted on a very large dataset consisting of Kannada, Devanagari, and English documents. The proposed method is effective, and results are promising.