An Efficient Character Recognition Technique Using K-Nearest Neighbor Classifier

摘要

Optical Character Recognition (OCR) Systems offers human machine interaction and are commonly used in several important applications. A lot of research has already been accomplished on the character recognition in different languages. This paper presents a technique for recognition of Printed text with noise using Optical Character Recognition (OCR). The main steps of this system are pre-processing of the text including converting the text image to black/white and remove the noise from the text image, segmentation of the text image to each character, Feature extraction using zoning-based technique and classification. The System is implemented using MATLAB 2016a software application program and is still under development. Noise is removed from all the text images. The quality of the input document is very important to achieve high accuracy. The system is able to recognize characters in different 50 images.