Optical Character Recognition Using Deep Learning Techniques for Printed and Handwritten Documents