Recognition Persian Handwritten Characters using Hough Transform
Neda Aashrafi Khuzani, Maryam Sadat Mahmoodi · International Journal of Engineering Trends and Technology · 2014
Abstract — Today, expansion of computers influences in the business world has forced organizations to think about using computers to process and manage large amounts of information. Artificial intelligence provides approaches and algorithms for image processing. Using this algorithm, the computer can have the ability to identify the various components of an image. The present study presents a method for detecting Persian handwritten characters using Hough transform. One of the main characteristics of the Hough transform is to detect the lines, circles and other shapes that have analytic relationships. Using this method along with documents image processing, data fields written in paper documents can be separately extracted and then use them for storage, processing, management and reporting. For instance, it considers discrete Persian handwritten characters as objects in a scene, and then attempts to diagnoses them using the Hough transform as well as neural network. According to the performed tests and obtained results, the accuracy of 70 % was yielded in detection of characters. Keywords—image processing, neural network, edge detection, skeletonizing, Hough transform. I.