Generating Arabic Letters using Generative Adversarial Networks (GANs)

Abstract

The Generative modeling became recently one of the most important eld in deep learning, and the Generative Adversarial Networks is the new research line in this eld, the GANs have proven their ability to generate a high resolution of images and achieved remarkable success for computer vision in general. by the same way, we assume that this tool will be able to generate accurate Arabic letters. The expected bene t of this work is creating a perception by using GANs and explore the potential of this deep learning powerful tool to serve the Arabic calligraphy. As an impact for further studies in the future, this work may provide an insight into the possibility of generating new kinds of Arabic fonts (calligraphy).