LSA64

该数据集名为LSA64，包含了来自阿根廷手语的64个不同手势的3200个视频，由10名受试者对每个手势进行5次重复表演录制而成。这些手势既有动词也有名词，且在不同的环境下执行，以增强现实感。为了便于手部追踪和分割，受试者佩戴了彩色手套。该数据集支持对手语识别的研究，并包含了关于动作、位置和手型的统计数据。其规模为3200个视频涵盖64个手势，任务是对手语进行自动识别。

The dataset named LSA64 consists of 3200 video samples of 64 distinct gestures from Argentine Sign Language, recorded from 10 subjects who each performed every gesture 5 times repetitively. These gestures cover both verbs and nouns, and were executed in diverse contexts to enhance real-world authenticity. To facilitate hand tracking and segmentation, all subjects wore colored gloves during the recording sessions. This dataset supports research on sign language recognition, and provides statistical data related to gesture motion, spatial position, and hand shape. With 3200 videos spanning 64 gesture categories, the core task supported by this dataset is automatic sign language recognition.