A Segmentation-driven Handwritten Uighur Word Recognition Algorithm Based on Feedback Structure

Abstract

Uighur script is cursive in both printed and handwritten forms. For offline handwritten Uighur word, this study proposes a new segmentation-driven recognition algorithm that combines feedback structure and grapheme analysis. Firstly, a handwritten Uighur word is over-segmented into a two-queue grapheme sequence using a MSAC (main segmentation and additional clustering) algorithm. Secondly, a feedback-based grapheme merging strategy is designed to provide the best segmented character sequence and obtain the word recognition result. Three feedback errors accordingly are defined, which are error of grapheme shape, error of character recognition and word matching error. A word recognition rate of 90.82% is obtained during experiments conducted with a database consisting of 11,500 samples.