Improved linear density technique for segmentation in Arabic handwritten text recognition

The challenge in handwriting recognition, especially in the segmentation process, took the researchers’ attention. These Arabic handwritten text processes are a challenging job because their characters are generally both cursive and unconstrained. In this paper, a new segmentation technique is proposed for solving the problem of Arabic handwritten scripts, called ILDT. The proposed technique’s main objective is to use the word image’s vertical linear density for clarifying character boundaries and districting between characters. In the proposed method, three pre-processing steps are applied: fill close and open holes (missing circle), remove punctuation to clarify the area of ligature points and avoid characters overlapping, and crop the word image to remove excess white space. The goal of filling close and open holes is to increase the character’s pixel density and then apply the vertical linear density. The proposed technique calculates the distance histogram of vertical linear, aiming to discover local minima points to precisely determine the segmentation points. Several experiments were conducted, including elapsed CPU times and accuracies values. All comparative techniques are examined on a local benchmark database. The proposed method (ILDT) got almost all the best segmentation and recognition accuracy compared with other comparative methods.