Skip to content

javierlorenzod/toolbox

 
 

Repository files navigation


KAIST Multispectral Pedestrian Detection Benchmark

By Soonmin Hwang, Jaesik Park, Namil Kim, Yukyung Choi, In So Kweon at RCV Lab. (KAIST) Website teaserImage

 

Dataset Download

[Readme]

[All, Video (35.9GB)]

[All, Annotation (48MB)]

 

Download links for part of our dataset

Train

Set 00 / Day / Campus / 5.92GB / 17,498 frames / 11,016 objects [videos] [annotations]

Set 01 / Day / Road / 2.82GB / 8,035 frames / 8,550 objects [videos] [annotations]

Set 02 / Day / Downtown / 3.08GB / 7,866 frames / 11,493 objects [videos] [annotations]

Set 03 / Night / Campus / 2.40GB / 6,668 frames / 7,418 objects [videos] [annotations]

Set 04 / Night / Road / 2.88GB / 7,200 frames / 17,579 objects [videos] [annotations]

Set 05 / Night / Downtown / 1.01GB / 2,920 frames / 4,655 objects [videos] [annotations]

 

Test

Set 06 / Day / Campus / 4.78GB / 12,988 frames / 12,086 objects [videos] [annotations]

Set 07 / Day / Road / 3.04GB / 8,141 frames / 4,225 objects [videos] [annotations]

Set 08 / Day / Downtown / 3.50GB / 8,050 frames / 23,309 objects [videos] [annotations]

Set 09 / Night / Campus / 1.38GB / 3,500 frames / 3,577 objects [videos] [annotations]

Set 10 / Night / Road / 3.75GB / 8,902 frames / 4,987 objects [videos] [annotations]

Set 11 / Night / Downtown / 1.33GB / 3,560 frames / 6,655 objects [videos] [annotations]

 

Or you can download extracted png files. (consisting of a few thousands png images)

train20 (1.6GB)

test20 (1.6GB)

 

Toolbox

This is an extended version of Piotr's Computer Vision Matlab Toolbox to deal with multispectral images.
Original toolbox is here

 

Citation

If you use our extended toolbox or dataset in your research, please consider citing:

@inproceedings{hwang2015multispectral,
	Author = {Soonmin Hwang and Jaesik Park and Namil Kim and Yukyung Choi and In So Kweon},
	Title = {Multispectral Pedestrian Detection: Benchmark Dataset and Baselines},
	Booktitle = {Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
	Year = {2015}
}

  To reproduce the results in CVPR '15 paper, run 'detector/acfDemoKAIST.m'.

About

KAIST Multispectral Pedestrian Detection Benchmark [CVPR '15]

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • MATLAB 86.0%
  • C++ 10.5%
  • C 3.3%
  • Other 0.2%