OFFLINE KANNADA HANDWRITTEN WORDS RECOGNITION USING DEEP LEARNING

OFFLINE KANNADA HANDWRITTEN WORDS RECOGNITION USING DEEP LEARNING

G Gangashree1, Dr. Lakshman Naika R2

PG Scholor1, Department of Electronics and Communication Engineering.

Associative Professor2, Department of Electronics and Communication Engineering.

University B.D.T College of Engineering, Davanagere-577004, Karnataka, India.

ABSTRACT
This paper proposes an efficient method for Offline Kannada handwritten word recognition system which uses image preprocessing techniques. Chars 51K dataset was used for experimentation of the work. Experiments were performed on handwritten Kannada Varanamala, Ottakshara, and Kaagunitha consisting of 25 handwritten characters in 637 classes. Kannada characters are used in our dataset. i.e Varanamala, Kaagunitha and Ottakshara. In this report we used CNN for training our model where it consists of different layers.

Keywords: A4 sheets, Black marker pen, Deep Learning, CNN.