Performance Comparison of SVM and K-NN for Oriya Character Recognition
Journal Title: International Journal of Advanced Computer Science & Applications - Year 2011, Vol 0, Issue 1
Abstract
Image classification is one of the most important branch of Artificial intelligence; its application seems to be in a promising direction in the development of character recognition in Optical Character Recognition (OCR). Character recognition (CR) has been extensively studied in the last half century and progressed to the level, sufficient to produce technology driven applications. Now the rapidly growing computational power enables the implementation of the present CR methodologies and also creates an increasing demand on many emerging application domains, which require more advanced methodologies. Researchers for the recognition of Indic Languages and scripts are comparatively less with other languages. There are lots of different machine learning algorithms used for image classification nowadays. In this paper, we discuss the characteristics of some classification methods such as Support Vector Machines (SVM) and K-Nearest Neighborhood (K-NN) that have been applied to Oriya characters. We will discuss the performance of each algorithm for character classification based on drawing their learning curve, selecting parameters and comparing their correct rate on different categories of Oriya characters. It has been observed that Support Vector Machines outperforms among both the classifiers.
Authors and Affiliations
Sanghamitra Mohanty , Himadri Nandini Das Bebartta
Keywords
Related Articles
- EP ID EP92328
- DOI -
- Views 121
- Downloads 0
An Image Encryption Technique based on Chaotic S-Box and Arnold Transform
In recent years, chaos has been extensively used in cryptographic systems. In this regard, one dimensional chaotic maps gained increased attention because of their intrinsic simplicity and ease in application. Many image...
Social Networks’ Benefits, Privacy, and Identity Theft: KSA Case Study
Privacy breaches and Identity Theft cases are increasing at an alarming rate. Social Networking Sites (SN’s) are making it worse. Facebook (FB), Twitter and other SN’s offer attackers a wide and easily accessible platfor...
Dynamic Programming Inspired Genetic Programming to Solve Regression Problems
The candidate solution in traditional Genetic Pro-graming is evolved through prescribed number of generations using fitness measure. It has been observed that, improvement of GP on different problems is insignificant at...
A Developed Collaborative Filtering Similarity Method to Improve the Accuracy of Recommendations under Data Sparsity
This paper presented a new similarity method to improve the accuracy of traditional Collaborative Filtering (CF) method under sparse data issue. CF provides the user with items, that what they need, based on analyses the...
Simulation of Building Evacuation: Performance Analysis and Simplified Model
Crowd evacuation from industrial buildings, factories, theatres, protest areas, festivals, exhibitions and religious/sports gatherings is very crucial in many dangerous scenarios, such as fire, earthquakes, threat and at...