DIOR
收藏资源简介:
DIOR数据集由纽约州立大学布法罗分校计算机科学与工程系创建,专注于室内外远距离3D/2D骨骼步态收集。该数据集包含14个主题和164.9万RGB帧,其中3D/2D骨骼步态标签包括20万帧来自远距离摄像机。DIOR利用先进的3D计算机视觉技术,在室内环境中实现像素级精度,并在室外远距离环境中采用低成本的混合3D计算机视觉和学习管道,仅使用4个低成本RGB摄像机即可实现精确的骨骼标注。此数据集旨在解决远距离步态识别问题,适用于活动识别、身份识别等领域,特别是在需要从屋顶摄像头、无人机摄像头等远距离识别个人的场景中具有重要应用价值。
The DIOR dataset was created by the Department of Computer Science and Engineering, University at Buffalo, The State University of New York. It focuses on collecting 3D/2D skeletal gait data over long distances in both indoor and outdoor environments. This dataset comprises 14 subjects and 1.649 million RGB frames, among which 200,000 frames with 3D/2D skeletal gait labels are captured by long-distance cameras. DIOR leverages advanced 3D computer vision technologies to achieve pixel-level accuracy in indoor environments, while adopting a low-cost hybrid 3D computer vision and learning pipeline for outdoor long-distance scenarios, enabling precise skeletal annotation with only four low-cost RGB cameras. This dataset aims to solve the problem of long-distance gait recognition and is applicable to fields such as activity recognition and identity recognition, with particularly important application value in scenarios requiring individual identification from long-distance cameras like roof-mounted cameras and drone cameras.

- 1Object Detection in Optical Remote Sensing Images: A Survey and A New Benchmark · 2019年



