Large-scale Multiview 3D Hand Pose Dataset
收藏资源简介:
本数据集名为‘大规模多视角3D手势数据集’,由阿利坎特大学计算机研究所创建。数据集包含超过20,500个标注的彩色手部图像,每张图像都配有不同类型的标注,如边界框和2D及3D关节位置。数据集通过定制的捕捉设备生成,该设备包括四个摄像头和一个Leap Motion控制器,用于提供高精度的手部跟踪。此数据集旨在支持深度学习技术在人机交互、用户界面和虚拟现实应用中的手势识别研究,解决现有数据集样本数量有限、标注抽象或仅包含深度图的问题。
This dataset, named 'Large-scale Multi-view 3D Gesture Dataset', was created by the Institute of Computer Science, University of Alicante. It contains over 20,500 annotated color hand images, each paired with multiple types of annotations including bounding boxes, 2D and 3D joint positions. The dataset is generated using a custom-built capture device that comprises four cameras and a Leap Motion controller to provide high-precision hand tracking. This dataset is intended to support research on gesture recognition leveraging deep learning technologies for human-computer interaction, user interface and virtual reality applications, addressing the limitations of existing datasets such as limited sample size, abstract annotations or only containing depth maps.



