Gujarati Handwritten Digit (GHD) Dataset

This article presents the Gujarati-HWD dataset, an open-source collection of handwritten Gujarati digits intended to facilitate research in machine learning, deep learning, and regional script recognition. The dataset comprises 11,150grayscale images of digits 0 to 9, collected from 360 participants across Gujarat, India, representing diverse age groups, genders, and educational backgrounds. Each sample was digitized at high resolution and preprocessed through grayscale conversion, normalization, noise reduction, and resizing to 128x128 pixels. Metadata such as participant age and gender accompanies the dataset, enabling demographic analysis.