Kannada Character Recognition Using Convolutional Neural Networks
Rakshith Shetty, M Sankalpa, Shradha Bhatia, Kaushik Mallibhat, Nalini C. Iyer · 2024
India is a country with many states, the form of documentation in Government offices and healthcare departments are in regional languages, and digitizing the document is a challenge. Thus, the current study focuses on Kannada (a regional language in Karnataka state, India) handwritten character identification approach uses techniques of image pre-processing techniques to improve image quality and deep learning approaches for feature extraction and thereby classification of Kannada characters. The Chars74K dataset was utilized to train the network. The trained model is capable of classifying handwritten Kannada vowels and consonants. In comparison to other architectures, MobileN etV2 has demonstrated high performance.