JRDB
收藏资源简介:
JRDB是由斯坦福大学视觉与学习实验室创建的一个新颖的自我中心数据集,专门用于研究在建筑环境中机器人的视觉感知。该数据集包含64分钟的标注多模态传感器数据,包括立体圆柱形360°RGB视频、3D点云、音频信号等。数据集涵盖了室内环境和行人区域等传统上代表性不足的场景,所有数据都是从机器人的自我视角收集的,无论是静止还是导航状态。JRDB数据集已被用于2D和3D人体检测和跟踪的基准测试,旨在为自我中心机器人视觉、自主导航和社交机器人在人类环境中的感知任务研究提供新的数据源和测试平台。
JRDB is a novel egocentric dataset created by the Stanford Vision and Learning Lab, specifically designed for research on robotic visual perception in built environments. This dataset contains 64 minutes of annotated multimodal sensor data, including stereoscopic cylindrical 360° RGB video, 3D point clouds, audio signals, and more. The dataset covers traditionally underrepresented scenarios such as indoor environments and pedestrian areas, with all data collected from the robot's egocentric perspective, whether stationary or navigating. The JRDB dataset has been used as a benchmark for 2D and 3D human detection and tracking, aiming to provide a new data source and testbed for research on perception tasks in egocentric robotic vision, autonomous navigation, and social robotics within human environments.

- 1JRDB: A Dataset and Benchmark of Egocentric Robot Visual Perception of Humans in Built Environments斯坦福大学视觉与学习实验室 · 2021年



