A supplementary material for "Bangla Sign Language Translator for Deaf and Speech Impaired People using Deep LSTM", accepted in International Journal of Speech Technology, Springer, 2025Authors: Aonmoy Das, Ananna Dev Aishi, Masbah Uddin Toha, Md. Fazlul Kader This dataset is designed for classifying 100 different sign language gestures using LSTM-based models. It consists of two .rar files: one containing raw video recordings, with each sign having 5 samples, organized into separate folders, and another containing numpy arrays of extracted keypoints from these videos. Each video is converted into 30 sequences of 30 frames, stored in a structured format that mirrors the video dataset.