BdSL36

BdSL36是由阿萨努多拉大学科学技术学院创建的大型数据集，专注于孟加拉国手语字母识别。该数据集包含超过四百万张图像，分为36个类别，旨在通过背景增强技术提高数据集的多样性。数据集的创建过程包括志愿者收集图像、专家审核和数据增强等步骤。BdSL36的应用领域主要在于开发实时手语字母识别系统，以帮助听力障碍人士与非手语使用者之间的沟通。

BdSL36 is a large-scale dataset developed by the School of Science and Technology of Asanudola University, dedicated to Bangladeshi Sign Language (BdSL) alphabet recognition. It contains over four million images categorized into 36 distinct classes, and is designed to boost dataset diversity via background augmentation techniques. The dataset creation process includes multiple steps such as volunteer-led image collection, expert review, and data augmentation. The primary application scenario of BdSL36 is the development of real-time sign language alphabet recognition systems, which aims to assist communication between hearing-impaired individuals and non-sign-language users.