Public Repository for Grantha Data Set

The dataset contains handwritten samples of Grantha numbers and vowels, totaling 44 different characters (10 numbers and 34 vowels). The data was collected on standard A4 sheets and scanned using a mobile phone. The handwritten samples were obtained from a diverse group of 150 subjects across various age groups. The data underwent segmentation, preprocessing, and was stored in a publicly accessible repository. After removing obscured images and scribbles, the final dataset comprises 5,852 digitized images, which include 1,330 Grantha numbers (133 samples for each number) and 4,522 vowels (133 samples for each vowel). This data is meticulously organized into respective folders. Additionally, the dataset is available in the form of 44 CSV (comma-separated values) files, with each file representing a unique character or number, and the corresponding labels attached.

本数据集包含格兰萨（Grantha）数字与元音的手写样本，共计44种不同字符（10个数字与34个元音）。数据采集于标准A4纸张，并通过手机完成扫描。手写样本来自150名覆盖不同年龄段的多元化受试群体。该数据经过分割、预处理流程后，存储于公开可访问的存储库中。在移除模糊图像与涂鸦内容后，最终数据集包含5852张数字化图像，其中1330张为格兰萨数字图像（每个数字对应133个样本），4522张为元音图像（每个元音对应133个样本）。本数据集已按类别细致规整至对应的专属文件夹中。此外，该数据集还提供了44个CSV（逗号分隔值，Comma-Separated Values）格式文件，每个文件对应一种独特的字符或数字，并附带相应的标签。