LSA-T

LSA-T是首个连续的阿根廷手语（LSA）数据集，由拉普拉塔国立大学计算机研究所LIDI创建。该数据集包含14,880个来自CN Sordos YouTube频道的句子级视频，每个视频都有标签和关键点注释。数据集的创建过程包括从视频中提取有效句子并进行关键点分析。LSA-T旨在通过提供丰富的手语视频资源，促进手语翻译（SLT）技术的发展，从而提高聋人群体的沟通和信息获取能力。LSA-T is the first continuous Argentine Sign Language (LSA) dataset, developed by LIDI, Institute of Computer Science, National University of La Plata. The dataset comprises 14,880 sentence-level videos sourced from the CN Sordos YouTube channel, with each video paired with labels and keypoint annotations. The dataset construction process includes extracting valid sentences from the videos and conducting keypoint analysis. LSA-T aims to promote the development of Sign Language Translation (SLT) technology by providing abundant sign language video resources, thereby improving the communication and information acquisition capabilities of the deaf community.

搜集汇总

背景与挑战

背景概述

LSA-T是首个连续的阿根廷手语（LSA）数据集，包含14,880个句子级视频，每个视频都有标签和关键点注释，源自CN Sordos YouTube频道。该数据集通过提供丰富的手语资源，旨在促进手语翻译技术发展，提升聋人群体的沟通和信息获取能力。