STM Journals Peer-reviewed research from Consortium e-Learning Network

Journal of Artificial Intelligence Research & Advances

Historical Kannada Handwritten Character Recognition using K-Nearest Neighbour Technique

  1. Chandrashekar Gudada
  2. Parashuram Bannigidada

Abstract

AbstractMost of the historical Kannada handwritten documents are preserved in the manuscript preservation center and archaeological department, and these documents are generally degraded in nature and it is very difficult to read and understand the contents in it. Hence, it is very much essential to digitize the historical Kannada handwritten document and recognize its originality of dynasty. The main objective of this paper is to digitize and recognize the historical Kannada handwritten documents by applying bonding box segmentation method and extracting the geometrical shape features, classification is performed by using K-nearest neighbor classifier. The average classification accuracy of the historical Kannada handwritten character from the different dynasties based on their age-type is: Kadamba 97.83%, Badami chalukya 97.78%, Kalyana chalukya 97.92%, Hoysala 97.87%, Vijayanagara 100%, Mysore wodeyars 97.87% and Aadhunika Kannada 97.96%. The results are compared with manual results obtained by the epigraphists and language experts, which demonstrate the efficacy of the proposed method.Keywords: Restoration, segmentation, Kannada; K-NN, character recognition, handwritten script, historical documents, document image analysisCite this ArticleParashuram Bannigidad, Chandrashekar Gudada. Historical Kannada Handwritten Character Recognition using K-Nearest Neighbour Technique. Journal of Artificial Intelligence Research & Advances. 2019; 6(1): 23–29p.

Keywords

Support