Convolutional Deep Learning Network for Handwritten Arabic Script Recognition

During the last years, deep convolution networks have emerged to become widespread, resulting in substantial gains in various benchmarks. In this paper, Convolutional Deep Belief Networks (CDBN) is applied to learn automatically the finest discriminative features from textual image data consisting of Arabic Handwritten Script. This architecture is able to lay hold of the advantages of Deep Belief Network and Convolutional Neural Network. We subjoin Regularization methods to our CDBN model so that we can address the issue of over-fitting. We evaluated our proposed model on high-level dimension in Arabic textual images. The obtained outcomes from the experiments prove that our model is more effective if compared to the ultra-modern results in handwritten script recognition using IFN/ENIT data sets.