Awesome

VisDrone2018-MOT Benchmark Toolkit

Introduction

This is the documentation of the VisDrone2018 competitions development kit for multiple object tracking (MOT) challenge.

This code library is for research purpose only, which is modified based on the toolkits in MOTChallenge [1] and PASCAL VOC [2].

The code is tested on the Windows 10 and macOS Sierra 10.12.6 systems, with the Matlab 2013a/2014b/2016b/2017b platforms.

If you have any questions, please contact us (email:tju.drone.vision@gmail.com).

Citation

If you use our toolkit or dataset, please cite our paper as follows:

@article{zhuvisdrone2018,

title={Vision Meets Drones: A Challenge},

author={Zhu, Pengfei and Wen, Longyin and Bian, Xiao and Haibin, Ling and Hu, Qinghua},

journal={arXiv preprint:1804.07437},

year={2018}

}

Dataset

For MOT competition, there are three sets of data and labels: training data, validation data, and test-challenge data. There is no overlap between the three sets.

                                                      Number of snippets
------------------------------------------------------------------------------------------------
 Dataset                           Training              Validation            Test-Challenge
------------------------------------------------------------------------------------------------
 Multiple object tracking           56 clips               7 clips                16 clips
                                  24,201 frames          2,819 frames           6,333 frames
------------------------------------------------------------------------------------------------

The challenge aims to recover the object trajectories with (Task 4B) or without (Task 4A) the detection results in each video frame. Notably, in the VisDrone2018 Challenge, we only consider five object categories in multi-object tracking, i.e., car, bus, truck, pedestrian, and van. We manually annotate the bounding boxes of different objects and ignored regiones in each video frame. Annotations on the training and validation sets are publicly available.

The dataset can be downloaded at

http://www.aiskyeye.com/

Evaluation Routines

The notes for the folders:

main functions
- evalMOT.m is the main function to evaluate your tracker
  - put the source codes in ./trackers/ (please refer the source codes of GOG tracker)
  - modify the dataset path and result path
  - use "isSeqDisplay" to show the groundtruth and detections
  - select the evaluated task, i.e, Task 4A without detection results, Task 4B with detection results
- evaluateTrackA.m is the main function to evaluate your tracker using the measures in Task 4A without detection results.
- evaluateTrackB.m is the main function to evaluate your tracker using the measures in Task 4B with the Faster RCNN detection results.

MOT Submission Format

Submission of the results will consist of TXT files with one line per predicted object.It looks as follows:

    <frame_index>,<target_id>,<bbox_left>,<bbox_top>,<bbox_width>,<bbox_height>,<score>,<object_category>,<truncation>,<occlusion>

 -----------------------------------------------------------------------------------------------------------------------------------
       Name	                                      Description
 -----------------------------------------------------------------------------------------------------------------------------------
   <frame_index>	  The frame index of the video frame
   
    <target_id>	          In the DETECTION result file, the identity of the target should be set to the constant -1.
		          In the GROUNDTRUTH file, the identity of the target is used to provide the temporal corresponding 
		          relation of the bounding boxes in different frames.
			  
    <bbox_left>	          The x coordinate of the top-left corner of the predicted bounding box

    <bbox_top>	          The y coordinate of the top-left corner of the predicted object bounding box

    <bbox_width>	  The width in pixels of the predicted object bounding box

    <bbox_height>	  The height in pixels of the predicted object bounding box

      <score>	          The score in the DETECTION file indicates the confidence of the predicted bounding box enclosing 
                          an object instance.
                          The score in GROUNDTRUTH file is set to 1 or 0. 1 indicates the bounding box is considered in evaluation, 
		          while 0 indicates the bounding box will be ignored.
			  
  <object_category>	  The object category indicates the type of annotated object, (i.e., ignored regions(0), pedestrian(1), 
                          people(2), bicycle(3), car(4), van(5), truck(6), tricycle(7), awning-tricycle(8), bus(9), motor(10), 
                          others(11))
		      
    <truncation>	  The score in the DETECTION file should be set to the constant -1.
                          The score in the GROUNDTRUTH file indicates the degree of object parts appears outside a frame 
		          (i.e., no truncation = 0 (truncation ratio 0%), and partial truncation = 1 (truncation ratio 1% ~ 50%)).
		      
     <occlusion>	  The score in the DETECTION file should be set to the constant -1.
                          The score in the GROUNDTRUTH file indicates the fraction of objects being occluded 
		          (i.e., no occlusion = 0 (occlusion ratio 0%), partial occlusion = 1 (occlusion ratio 1% ~ 50%), 
		          and heavy occlusion = 2 (occlusion ratio 50% ~ 100%)).

The sample submission of the GOG tracker can be found in our website.

References

[1] A. Milan, L. Leal-Taixe, K. Schindler, D. Cremers, S. Roth, and I. Reid, "Multiple Object Tracking Benchmark 2016", https://motchallenge.net/results/MOT16/.

[2] E. Park, W. Liu, O. Russakovsky, J. Deng, F.-F. Li, and A. Berg, "Large Scale Visual Recognition Challenge 2017", http://imagenet.org/challenges/LSVRC/2017

Version History 1.0.2 - Jul 1, 2019

fix the bugs in droping objects in ignored regions

1.0.1 - May 2, 2018

fix the bugs in evaluating each object category
improve the evaluation speed

1.0.0 - Apr 27, 2018

initial release