TOFFE dataset
收藏资源简介:
TOFFE数据集是由普渡大学埃尔莫尔家族电气与计算机工程学院开发的一个合成事件视觉数据集,旨在支持高速运动场景下的物体检测与跟踪任务。该数据集在Gazebo模拟器中生成,包含多种传感器数据,如帧图像、深度信息、事件数据和6自由度姿态信息,采样率高达20000次/秒。数据集提供了精确的物体姿态和速度地面真值,适用于训练和评估高速运动场景下的算法。TOFFE数据集通过模拟不同速度和轨迹的物体运动,为研究提供了丰富的高动态范围数据,特别适合用于边缘计算和低功耗场景下的实时处理任务。
The TOFFE dataset is a synthetic event-based vision dataset developed by the Elmore Family School of Electrical and Computer Engineering at Purdue University, designed to support object detection and tracking tasks in high-speed motion scenarios. Generated using the Gazebo simulator, this dataset includes multiple types of sensor data such as frame images, depth information, event data, and 6-degree-of-freedom (6DoF) pose information, with a sampling rate of up to 20,000 samples per second. The dataset provides accurate ground truth for object pose and velocity, which is suitable for training and evaluating algorithms for high-speed motion scenarios. By simulating object motions with varying speeds and trajectories, the TOFFE dataset offers rich high dynamic range data for research, and is particularly well-suited for real-time processing tasks in edge computing and low-power scenarios.




