Self-Learning Based Cognitive Reading and Character Recognition in Image Processing Techniques

Abstract

There are many techniques involved in Handwritten character recognition and many of the methods uses common process like pre-processing, segmentation, stroke identification and character interpretation. But the technique applied for these steps differ from different implementation. Building the training data is a tedious task but it is more important as the success of entire recognition system lies on the amount of training the neural network and building the knowledge base. Hence this is an important task and it is costly due to amount of time and resource used for training the dataset. This article explains some of the methods in Cognitive reading in image processing for character recognition and introduces a self-learning based training system which provides improved method in less time/resource consuming in automatically training the knowledge base.