top of page

KAIST Multispectral Pedestrian Detection Benchmark


We developed imaging hardware consisting of a color camera, a thermal camera and a beam splitter to capture the aligned multispectral (RGB color + Thermal) images. With this hardware, we captured various regular traffic scenes at day and night time to consider changes in light conditions.

The KAIST Multispectral Pedestrian Dataset consists of 95k color-thermal pairs (640x480, 20Hz) taken from a vehicle. All the pairs are manually annotated (person, people, cyclist) for the total of 103,128 dense annotations and 1,182 unique pedestrians. The annotation includes temporal correspondence between bounding boxes like Caltech Pedestrian Dataset. More infomation can be found in our CVPR 2015 paper.

Download

  • KAIST Multispectral Pedestrian Dataset. (Click to download)

  • Data Format. (Compatible with Caltech Pedestrian Dataset Format)

  • Extended Video Annotation Tool for Multispectral Images.

Contact

Please contact Soonmin Hwang [smhwang at rcv.kaist.ac.kr] with questions or comments.

References

Soonmin Hwang, Jaesik Park, Namil Kim, Yukyung Choi and In So Kweon,

Multispectral Pedestrian Detection: Benchmark Dataset and Baseline,

CVPR, 2015. [pdf|Ext. Abstract]

Change Log

2015.06.03. The "Multispectral Pedestrian Detection Benchmark" webpages will be opened.

© 2015 by Yukyung Choi. Proudly created withwith Wix.com

Korea Advanced Institute of 

Science and Technology

Robotics and Computer Vision

Laboratory

bottom of page