RSCNN: multi Correlation filter tracking using hierarchy of convolution features + minimum-cost network flow.

TUD-Crossing


Benchmark:

Short name:

RSCNN

Detector:

Public

Description:

n/a

Reference:

Heba Mahgoub, Khaled Mostafa, Khaled T. Wassif and Ibrahim Farag, “Multi-Target Tracking Using Hierarchical Convolutional Features and Motion Cues” International Journal of Advanced Computer Science and Applications(IJACSA), 8(11), 2017.

Last submitted:

October 07, 2017 (2 years ago)

Published:

July 08, 2017 at 23:21:06 CET

Submissions:

4

Project page / code:

n/a

Open source:

No

Hardware:

Runtime:

4.0 Hz

Benchmark performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
2D MOT 201529.537.073.193 (12.9)262 (36.3)11,86630,47450.472.32.1976 (19.4)1,176 (23.3)

Detailed performance:

Sequence MOTA IDF1 MOTP MT ML FP FN Recall Precision FAF ID Sw. Frag
ADL-Rundle-120.133.172.6833,4573,82858.961.36.9150152
ADL-Rundle-336.039.177.5781,2505,17149.180.02.08681
AVG-TownCentre28.640.169.720797484,15741.880.01.7200283
ETH-Crossing26.930.477.91182270629.693.10.155
ETH-Jelmoli51.157.676.1141433986965.783.10.83341
ETH-Linthescher35.440.476.8171161435,57237.695.90.15564
KITTI-1641.342.072.02131164662.077.21.54257
KITTI-1930.844.567.37141,0252,57951.772.91.095160
PETS09-S2L240.029.371.8466784,88549.387.51.6222255
TUD-Crossing75.563.876.0804320981.095.40.21818
Venice-1-26.523.770.2533,8501,85259.441.38.67060

Raw data: