Handwritten Character Recognition Based on the Specificity and the Singularity of the Arabic Language

Abstract

-A good Arabic handwritten recognition system must consider the characteristics of Arabic letters which can be explicit such as the presence of diacritics or implicit such as the baseline information (a virtual line on which cursive text are aligned and/join). In order to find an adequate method of features extraction, we have taken into consideration the nature of the Arabic characters. The paper investigate two methods based on two different visions: one describes the image in terms of the distribution of pixels, and the other describes it in terms of local patterns. Spatial Distribution of Pixels (SDP) is used according to the first vision; whereas Local Binary Patterns (LBP) are used for the second one. Tested on the Arabic portion of the Isolated Farsi Handwritten Character Database (IFHCDB) and using neural networks as a classifier, SDP achieve a recognition rate around 94% while LBP achieve a recognition rate of about 96%.

Spatial and Textural Aspects for Arabic Handwritten Characters Recognition

“…6. As in [1], the use of LBP histogram extracted from the smallest rectangle containing the character and its diacritical points with a feedforward neural network and a Scaled conjugate Gradient as training algorithm gives a recognition rate of 88.47% which is reasonable, since in the computation of LBP Histogram the spatial information is lost.…”

“…So in order to integrate the spatial aspects, as in [1], we simply split the bounding box of the character into four regions from the center point between the centroid of the character and the centroid of diacritics and extract the histogram from each region and then concatenates them to construct the feature vector (Fig. 7).…”

“…The present paper is an extension of our work in [1]. We will first compare the recognition accuracy of two feature extraction methods respecting the spatial and textural aspects, namely modified Bitmap Sampling method and Local Binary Patterns (LBP).…”

Segmentation of Arabic Handwritten Documents into Text Lines using Watershed Transform

“…In this paper we focus on the text line segmentation which could be defined according to [1] in the process of assigning the same label to units that are partially aligned. We are interested in Arabic handwritten documents which are more challenging than Latin documents and that is mainly due to the semi-cursive nature of the Arabic script which is characterized by its calligraphy and the presence of ascending and descending character, the overlapping between piece of Arabic words and also the diacritical points located either above or below characters [24].…”

IJIMAI Editor's Note - Vol. 4 Issue 4

“…Boulid, Y. et al [6] write about handwritten character recognition. A good Arabic handwritten recognition system must consider the characteristics of Arabic letters which can be explicit such as the presence of diacritics or implicit such as the baseline information (a virtual line on which cursive text are aligned and/join).…”