Siamese Neural Networks for Kannada Handwritten Dataset
S Tejas, Kusumika Krori Dutta · 2022 IEEE 3rd Global Conference for Advancement in Technology (GCAT) · 2022
With the increasing usage of Artificial intelligence (AI), many research challenges can be easily addressed and one such challenge is Handwritten character recognition. Kannada is one of the south Indian languages, eighth most spoken language in India. Many of the alphabets in Kannada are very similar and handwritten script classification poses a big challenge. Lack of proper dataset unlike other languages like English, Hindi, Bengali, etc., restricts good research in this field. This paper deals with creation to classification of 603 classes of Kannada handwritten character with optimized dataset using Siamese Neural Net (SNN), as this technique helps to classify with minimal training set and with better accuracy. This paper aims to analyze the performance of Siamese neural network in Handwritten Kannada character recognition to enhance the research work in various other languages.