In the last few years, deep learning-based models have made significant inroads into the field of handwriting recognition. However, deep learning requires the availability of massive labelled data and considerable computation for training or automatic feature extraction. The role of handcrafted features and their significance is still crucial for a specific language type because it is a unique way of writing the characters. These are primitive segments that describe the letter horizontally or vertically distinguish an Arabic letter. This article develops a new type of feature for handwriting using Segments Interpolation (SI) to find the best fitting line in each of the windows and build a model for finding the best operating point window size for SI features. The experimental design was done on two subsets of the Institute for Communications Technology/Ecole Nationale d'Ingénieurs de Tunis (IFN/ENIT) database. The first one contains 10 classes (C10), and the second one has 22 classes (C22). The extracted features were trained with Support Vector Machine (SVM) and Extreme Learning Machine (ELM) with different kernels and activation functions. The evaluation metrics from a classification perspective (Accuracy, Precision, Recall and Fmeasure) were applied. As a result, SI shows significant results with SVM 90.10% accuracy for C10 and 88.53% accuracy for C22.INDEX TERMS Arabic handwriting word recognition, classification, ELM, feature extraction, Segments Interpolation and SVM.
I. BACKGROUNDHandwriting recognition is a dynamic model and simulation environment that is considered a part of pattern recognition. It can contribute an essential benefit to our real life [1]. The diversity of handwriting recognition comes with extensive usage of a massive number of costly computational aspects. Currently, the technology provides an exceptionally smooth technique and, at the same time, hides the bright side of handwriting text. Several applications where handwriting recognition is necessary, such as bank cheques [2], postal addresses [3], and handwritten form processing [4].Numerous studies on handwriting recognition, especially for the Latin script [2][3], have been conducted over the last few decades. There are quite good results for machine printed text recognition with over 99% accuracy, for instance. However, in Arabic handwriting recognition as against Latin [7], only minor studies have been carried out. Due to the intricacy of Arabic text and insufficient databases about this language [8]. Recognition of Arabic text is in the initial phases compared to the recognition of Chinese, Latin, and Japanese manuscripts. Furthermore, there are several challenges in Arabic writing recognition practices from the data's cursive form. These challenges arise due to several aspects, like the Arabic writing setup that is cursive, the pen, the writing style, and other elements.Arabic handwriting recognition consists of two classes, online and offline [9], respectively. First, the characteristics of This work is licensed under a Creativ...
