dimi-fn / kannada-mnist-digits

Classification & Prediction of Images of Handwritten Digits in the Kannada Language.

*Note:

Due to a GitHub bug (issue #3035 & #3555), sometimes the notebook files (files ending in ".ipynb") may not render. Please reload the page until the content can be displayed, or click here to view the shared Google Colab notebook file.

Kannada - Mnist

This is a Machine Learning project using Deep Neural Networks (DNNs) and Convolutional Neural Networks (CNNs) with Tensorflow and Keras. More specifically, it is about a classification problem in regard to the recognition and prediction of the Kannada MNIST hand-written digits.

Source

The project uses part of data that was published in: Prabhu, Vinay Uday. "Kannada-MNIST: A new handwritten digits dataset for the Kannada language." arXiv preprint arXiv:1908.01242 (2019).

It is about an alternative dataset with reference to the popular MNIST digits database, and the project was constructed in the context of a kaggle class competition (module CS985 Machine Learning) at Strathclyde University (academic year 2019-20).

Description

"Kannada" is a language spoken predominantly by people of Karnataka in southwestern India. The language has roughly 45 million native speakers and is written using the Kannada script.

The format is similar to MNIST in terms of how the data is structured. Each image is 28 pixels in height and 28 pixels in width, for a total of 784 pixels in total. Each pixel has a single pixel-value associated with it, indicating the lightness or darkness of that pixel, with higher numbers meaning darker. This pixel-value is an integer between 0 and 255, inclusive, and each pixel column in the training set has a name like pixel{x}, where x is an integer between 0 and 783, inclusive. Overall, there are 10 classes since the handwritten digits range from 0 to 9.

Purpose

The classification of the images of hand-written digits, and the prediction of those numbers. A range of deep neural network architectures and techniques were applied with the final goal of finding the optimal model for the digits' predictions.

Environment

Google colab with GPU