Method

Hierarchical MOT Using Convolutional Features with Occlusiong Handling [on] [CCF-MOT]
[Anonymous Submission]

Submitted on 5 Mar. 2016 18:40 by
[Anonymous Submission]

Running time:1.1 s
Environment:1 core @ 3.6 Ghz (MATLAB)

Method Description:
use Regionlet detetions available at:
http://www.cvlibs.net/datasets/kitti/eval_tracking.p
hp
Parameters:
Learned from training data and applied on the
testing
data
Latex Bibtex:

Detailed Results

From all 29 test sequences, our benchmark computes the commonly used tracking metrics CLEARMOT, MT/PT/ML, identity switches, and fragmentations [1,2]. The tables below show all of these metrics.


Benchmark MOTA MOTP MODA MODP
CAR 77.08 % 78.36 % 77.28 % 82.96 %
PEDESTRIAN 44.52 % 68.38 % 45.43 % 91.61 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 82.05 % 97.02 % 88.91 % 31317 961 6853 8.64 % 34818 779
PEDESTRIAN 54.24 % 86.53 % 66.68 % 12641 1967 10666 17.68 % 16624 245

Benchmark MT PT ML IDS FRAG
CAR 52.62 % 34.31 % 13.08 % 69 391
PEDESTRIAN 24.40 % 38.49 % 37.11 % 211 976

This table as LaTeX


[1] K. Bernardin, R. Stiefelhagen: Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics. JIVP 2008.
[2] Y. Li, C. Huang, R. Nevatia: Learning to associate: HybridBoosted multi-target tracker for crowded scene. CVPR 2009.


eXTReMe Tracker