The Multiple Object Tracking Benchmark
Detection File
# detection 2D file is
1,-1,463,433,73.864,167.61,71.405,-1,-1,-1
1,-1,807,419,30.903,70.127,69.018,-1,-1,-1
1,-1,1474,407,79.918,181.35,42.996,-1,-1,-1
1,-1,399,429,26,59,31.164,-1,-1,-1
1,-1,1173,410,43.624,98.993,30.971,-1,-1,-1
1,-1,1363,394,79.918,181.35,29.228,-1,-1,-1
- 1, The first number indicates in which frame the object appears,
- 2, second number identifies that object as belonging to a trajectory by assigning a unique ID (set to −1 in a detection file, as no ID is assigned yet).
- Each object can be assigned to only one trajectory.
- 식별 ID, -1이면 ID가 아직 할당되지 않은것, Detection파일에서는 모두 -1인듯 Annotation파일에서는 ID존재
- 3~6, The next four numbers indicate the position of the bounding box of the pedestrian in 2D image coordinates.
- The position is indicated by the top-left corner as well as width and height of the bounding box.
7, This is followed by a single number, which in case of detections denotes their confidence score.
8~10, The last three numbers indicate the 3D position in real-world coordinates of the pedestrian.
- This position represents the feet of the person.
- In the case of 2D tracking, these values will be ignored and can be left at −1
Annotation File
# An example of such an annotation 2D file is:
1, 1, 794.2, 47.5, 71.2, 174.8, 1, -1, -1, -1
1, 2, 164.1, 19.6, 66.5, 163.2, 1, -1, -1, -1
1, 3, 875.4, 39.9, 25.3, 35.0, 0, -1, -1, -1
2, 1, 781.7, 25.1, 69.2, 170.2, 1, -1, -1, -1
- the 7th value (confidence score) acts as a flag whether the entry is to be considered.
- A value of 0 means that this particular instance is ignored in the evaluation,
- while a value of 1 is used to mark it as active.
MOTChallenge 2015: Towards a Benchmark for Multi-Target Tracking
Welcome to the UA-DETRAC Benchmark Suite!
TrackingNet: A Large-Scale Dataset and Benchmark for Object Tracking in the Wild, ECCV2018