Comparative Analysis for Gurmukhi and Devanagari Script at Word Level
Abstract
This paper presents a comparative analysis for Gurmukhi OCR with Devanagari OCR at word level. As OCR works in different stages and each stage has its own importance. In this paper, feature extraction and classification methods are discussed. Features are the set of minimum values of images to describe it uniquely. Before the classification, Features have been extracted from word images. In this paper, two feature extraction techniques Discrete Cosine Transform (DCT) and Gabor filter has been used. Gabor produces 189 features and DCT produces 100 features in zig-zag method. For training and testing, 50 different classes with 30-35 samples of each class for training and 10-15 samples for testing have been taken in Gurmukhi Script as well as Devanagari Scripts. For classification k-NN classifier with value of k=3, 5 and SVM have been used for performance comparison. Keyword: Feature extraction, Gabor Filter, DCT, k-NN Classifier, OCR.
Downloads
Published
How to Cite
Issue
Section
License
International Journal of Engineering Technology and Computer Research (IJETCR) by Articles is licensed under a Creative Commons Attribution 4.0 International License.