Implementation of Canny Edge Detection for Identification of Indonesian and Medina Usmani Khat Using Backpropagation

The Arabic writing on Al-Qur’an has variations of writing in the world that has been published. Since it has a variety of writing, it has differences how to writing it. The problem is how computer to recognizing that differences styles in the same word. In this research, we can recognize 2 styles Arabic writing to Medina and Indonesia styles using Canny edge detection for recognition pattern and backpropagation using for the identification of those styles. The results of recognizing images are 92 images for 100 images by 46 images Indonesian Khat and 46 Medina Khat.