Arabic handwritten digits recognition based on convolutional neural networks with resnet-34 model

Abstract

<span>Handwritten digits recognition has attracted the attention of researchers in pattern recognition fields, due to its importance in many applications in public real life, such as read bank checks and formal documents which is a continuous challenge in the last years. For this motivation, the researchers created several algorithms in recognition of different human languages, but the problem of the Arabic language is still widespread. Concerning its importance in many Arab and Islamic countries, because the people of these countries speak this language, However, there is still a little work to recognize patterns of letters and digits. In this paper, a new method is proposed that used pre-trained convolutional neural networks with resnet-34 model what is known as transfer learning for recognizing digits in the arabic language that provides us a high accuracy when this type of network is applied. This work uses a famous arabic handwritten digits dataset that called MADBase that contains 60000 training and 1000 testing samples that in later steps was converted to grayscale samples for convenient handling during the training process. This proposed method recorded the highest accuracy compared to previous methods, which is 99.6%.</span>

KurdSet Handwritten Digits Recognition Based on Different Convolutional Neural Networks Models

“…For instance, the state of art in [27] and [30] achieved accuracies of 94.30% and 97.40%, respectively. Furthermore, in [24] the accuracy result is 98.59% using RBM-CNN method, whereas in [29], the ResNet-34 model attained an accuracy of 99.60%. Table 5 shows the outcomes of the proposed method with the state of art approaches.…”

“…However, CNN on the local dataset alone achieved 80% accuracy. In a separate study by Savita et al [26] R. H. Finjan [29] in his article underscores the importance of recognizing Arabic numerals and recommends the utilization of (CNNs) employing the ResNet-34 architecture. The research involved a dataset comprising 60,000 Arabic handwritten digits, with 1000 testing samples converted to grayscale to facilitate training.…”

“…Dataset Accuracy RBM-CNN [24] AHDBase 98.59% Hybrid CNN [27] MNIST+AHDBase 94.30% ResNet-34 [29] AHDBase 99.60% CNN [30] AHDBase 97.40% Proposed Custom CNN AHDBase 99.65% Proposed ResNet50…”

Handwritten Javanese script recognition method based 12-layers deep convolutional neural network and data augmentation

“…Many recognition methods have been proposed. Some are used for Javanese script recognition [1]- [4], as well as non-Latin languages, such as Arabic [5]- [7], Tamil [8], Bangla or Bengali [9]- [11], Kannada [12], Gurmukhi [13], Tifinagh [14], and Thai [15]. Non-Latin character recognition is usually more difficult due to limited research and datasets and the relatively complex shapes of the character.…”

Characters recognition using keys points and convolutional neural network

“…Boutaounte et al [10] presented an approach based on characters decomposition into multiples geometric shapes (segment, arc) by detecting the points of branch and end and using a comparison between different methods of classification such as neural network (NN), k-mean and support vector machines (SVM). Other work such as Wardani, [11] and Aharrane et al [12] are based in the power of the convolutional neural network to build the OCR system, so the application of the convolutional neural network (CNN) or the deep learning in general are direct as other work in other language as the work of Finjan et al [13] or El-Sawy et al [14] applied in arabic language, Kadir et al, [15] used the CNN for number recognition in comparison with bag of features, Sadouk et al [16] used a handwritten database but the recognition rate were between 95% and 98% but Benaddy et al [17], used the CNN as a feature extractor and achieve a rate of 99%.…”