Text Segmentation of Historical Arabic Handwritten Manuscripts Using Projection Profile

摘要

The research in the field of recognizing Arabic handwritten texts is flourishing with the passage of time. Despite that, progress in this field is considered insignificant and has many challenges, one of which is that the available Arabic datasets are limited. Also, due to the complexity of Arabic language, it is difficult to perform necessary preprocessing steps, such as lines and characters segmentation, especially characters segmentation since Arabic letters are usually overlapped into each other. In this study, we perform lines and characters segmentation based on Projection Profile methods (PP) which will be applied on a historical Arabic manuscript that was collected from Umm Al-Qura University’s King Abdullah Bin Abdulaziz Library, Manuscripts department. The manuscript is a Moroccan Quranic manuscript that was written before more than 250 years ago. The results on lines segmentation are considered promising. On the other hand, poor results were achieved on characters segmentation due to the problem of complexity that was mentioned earlier.