Textural Features and Feature Selection in KAN-NADA Character Recognition
B N Ajay1, C Naveena2
1B N Ajay, Department of Electronics and Communications Engineering, Employee, HCL Technologies Tiruvallur, Tamil Nadu, India.
2C Naveena, Department of Electronics and Communications Engineering, Employee, HCL Technologies Tiruvallur, Tamil Nadu, India.
Manuscript received on 10 April 2019 | Revised Manuscript received on 17 April 2019 | Manuscript Published on 24 May 2019 | PP: 140-148 | Volume-8 Issue-6S3 April 2019 | Retrieval Number: F22240486S219/19©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: This paper describes an OCR system for printed text documents in Kannada, a South Indian language. Many commercial OCR systems are now available in the market, but most of these systems work for Roman, Chinese, Japanese and Arabic characters. There are no sufficient number of works on Indian language character recognition especially Kannada. In this work we proposed kannada character recognition system using texture features. Here we fuse the texture features like Local Binary Pattern, Local Binary Pattern Variance, Gray Level Local Texture Pattern, Gabor Filter Response and Wavelet Decomposition using concatenation rule and select discriminative texture features by employing wrapper feature selection methods. Finally, K-NN classifier is explored for the purpose of classification. In addition, we also explore the K-NN classifier with different distance functions. This method is simple to implement and realize, also it is computationally efficient.
Keywords: LTP, GLTP, KNN, SFS, SFFS.
Scope of the Article: Communications