Omni-MOT Dataset

The Omni-MOT is realistic CARLA based large-scale dataset with over 14M frames for multiple vehicle tracking . The dataset comprises 14M+ frames, 250K tracks, 110 million bounding boxes, three weather conditions, three crowd levels and three camera views in five simulated towns.

Download

[baidu], [google drive is not enough], [dropbox is also not that enough]

Structure

The published dataset comprises five-level structures. At the 1st level, we divide this dataset into the training set and the testing set. At the 2nd level, we employ five towns in the CARLA Simulator. At the 3rd level, we put the camera at different viewpoints containing five difficulty ranks. At the 4th level, we simulate different weathers. At the 5th level, we create a different number of cars.

Ground Truth

Index	Name	Description
0	frame index	0-based frame index
1	vehicle id	the unique id of vehicle (0-based)
2	bbox	represents left, top, right, bottom of the vehicle’s bounding box
6	3d bbox	The 8 points of the vehicle’s 3D bounding boxes in the image coordinate
14	vehicle position	the world coordinate of the vehicle’s center in the world coordinate
17	integrity	the integrity of the vehicle, value is in (0, 1)
18	velocity vector	the velocity vector in the world coordinate.
21	acceleration vector	the acceleration vector in the world coordinate.
24	wheel number	number of wheels
25	camera view size	the width and the height of the camera
27	camera FOV	the field of view of the camera
28	camera position	the camera coordinate in the world coordinate
31	camera rotation	the rotation of the camera
34	weather condition	the weather condition of current frame.

Sample Videos

Cite

@inproceedings{ShiJie20,
  author = {Shijie Sun, Naveed Aktar, XiangYu Song, Huansheng Song, Ajmal Mian, Mubarak Shah},
  title = {Simultaneous Detection and Tracking with Motion Modelling for Multiple Object Tracking},
  booktitle = {Proceedings of the European conference on computer vision (ECCV)}},
  year = {2020}

@inproceedings{Dosovitskiy17,
  title = {{CARLA}: {An} Open Urban Driving Simulator},
  author = {Alexey Dosovitskiy and German Ros and Felipe Codevilla and Antonio Lopez and Vladlen Koltun},
  booktitle = {Proceedings of the 1st Annual Conference on Robot Learning},
  pages = {1--16},
  year = {2017}
}

Copyright

CARLA specific assets are distributed under CC-BY License.
This dataset is opened and is released under the MIT License. Anyone can extend it. If you your work is based on it, please put the following texts in your description.
The methods provided on this page are published under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License . This means that you must attribute the work in the manner specified by the authors, you may not use this work for commercial purposes and if you alter, transform, or build upon this work, you may distribute the resulting work only under the same license. If you are interested in commercial usage you can contact us for further options.

shijieS / OmniMOTDataset