TS-RGBD数据集

TS-RGBD数据集是由USTHB大学的RIIMA实验室创建，专门为视觉障碍人士设计的剧场场景描述数据集。该数据集包含RGB、深度和骨骼序列三种数据类型，由Microsoft Kinect 1捕捉。数据集旨在通过图像字幕和人类动作识别技术，帮助视觉障碍人士理解和参与剧场表演。数据集包括36种剧场动作类别和38个剧场场景剧本，总计610个序列，平均每序列373帧。数据集的创建过程涉及与国家剧院的合作，以及在大学礼堂的实际拍摄。该数据集的应用领域主要集中在提升视觉障碍人士的剧场体验，通过技术手段增强他们对剧场场景的理解和参与。

The TS-RGBD dataset was developed by the RIIMA Laboratory of USTHB University, and it is a theater scene description dataset specifically designed for visually impaired individuals. This dataset encompasses three types of data: RGB, depth, and skeletal sequences, which were captured using Microsoft Kinect 1. The dataset aims to assist visually impaired people in understanding and participating in theater performances via image captioning and human action recognition technologies. The dataset includes 36 theater action categories and 38 theater scene scripts, with a total of 610 sequences, averaging 373 frames per sequence. The creation of this dataset involved cooperation with the National Theater and on-site filming in the university auditorium. The primary application areas of this dataset focus on enhancing the theater experience for visually impaired individuals, and strengthening their understanding of and engagement with theater scenes through technical means.