A Modified Back propagation Algorithm for Optical Character Recognition

Character Recognition (CR) has been an active area of research and due to its diverse applicable environment; it continues to be a challenging research topic. There is a clear need for optical character recognition in order to provide a fast and accurate method to search both existing images as well as large archives of existing paper documents. However, existing optical character recognition programs suffer from a flawed tradeoff between speed and accuracy, making it less attractive for large quantities of documents. In this thesis, we present a new neural network based method for optical character recognition as well as handwritten character recognition. Experimental results show that our proposed method achieves highest percent accuracy in optical character recognition. We present an overview of existing handwritten character recognition techniques. All these algorithms are described more or less on their own. Handwritten character recognition is a very popular and computationally expensive task. We also explain the fundamentals of handwritten character recognition. We describe today’s approaches for handwritten character recognition. From the broad variety of efficient techniques that have been developed we will compare the most important ones. We will systematize the techniques and analyze their performance based on both their run time performance and theoretical considerations. Their strengths and weaknesses are also investigated. It turns out that the behavior of the algorithms is much more similar as to be expected.