Kannada Character Recognition Using Convolutional Neural Network
Character recognition of reading handwritten alphabets in regional languages is a complex problem and needs machine learning for solving the problem effectively. There are multiple machine learning algorithms proposed to read the characters and deep learning approach, Convolutional Neural Network is one among them. In the work proposed we build a character recognition system for recognizing kannada alphabets. We have trained our model using our own dataset and trained model was tested with the test image. The model achieves 93% accuracy in recognizing kannada alphabets.
Keywords: Optical Character Recognition(OCR) Convolutional Neural Network, K-Nearest Neighbour, Linear Discriminant Analysis, Rectified Linear Unit