Proceedings ACS/IEEE International Conference on Computer Systems and Applications
DOI: 10.1109/aiccsa.2001.933960
Full text: Download
The segmentation and recognition of Arabic handwritten text has been an area of great interest in the past few years. However, only a small number of research papers and reports have been published in this area, due to the difficult problems associated with Arabic handwritten text processing. In this work, a technique is presented that segments handwritten Arabic text. A conventional algorithm is used for the initial segmentation of the text into connected blocks of characters. The algorithm then generates pre-segmentation points for these blocks. A neural network is subsequently used to verify the accuracy of these segmentation points. Two major problems were encountered. First, although the segmentation phase proved to be successful in the vertical segmentation of connected blocks of characters, it couldn't segment characters that were overlapping horizontally. Second, segmentation of handwritten Arabic text depends largely on contextual information, and not just on topographic features extracted from the characters