Pedestrian proposal generation using depth-aware scale estimation

Kihong Park, Seungryong Kim, Kwanghoon Sohn

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this work, we propose an efficient method that generates pedestrian proposals suitable for the autonomous vehicle. Our main intuition is that depth information provides an important cue to assign the scale of pedestrian proposals. Based on the observation that in a 3-D world coordinate the scales of pedestrians are almost similar, we formulate the scales of pedestrian patches by projecting 3-D models to an image plane with its corresponding depth. We also introduce a scale-aware binary description using both color and depth images. By using this descriptor, the regression models are trained to rank the pedestrian proposal candidates and adjust the proposal bounding boxes for an accurate localization. Our algorithm achieves significant performance gains compared to conventional proposal generation methods on the challenging KITTI dataset.

Original languageEnglish
Title of host publication2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings
PublisherIEEE Computer Society
Pages2045-2049
Number of pages5
ISBN (Electronic)9781509021758
DOIs
Publication statusPublished - 2018 Feb 20
Event24th IEEE International Conference on Image Processing, ICIP 2017 - Beijing, China
Duration: 2017 Sep 172017 Sep 20

Publication series

NameProceedings - International Conference on Image Processing, ICIP
Volume2017-September
ISSN (Print)1522-4880

Other

Other24th IEEE International Conference on Image Processing, ICIP 2017
CountryChina
CityBeijing
Period17/9/1717/9/20

Fingerprint

Color

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Park, K., Kim, S., & Sohn, K. (2018). Pedestrian proposal generation using depth-aware scale estimation. In 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings (pp. 2045-2049). (Proceedings - International Conference on Image Processing, ICIP; Vol. 2017-September). IEEE Computer Society. https://doi.org/10.1109/ICIP.2017.8296641
Park, Kihong ; Kim, Seungryong ; Sohn, Kwanghoon. / Pedestrian proposal generation using depth-aware scale estimation. 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE Computer Society, 2018. pp. 2045-2049 (Proceedings - International Conference on Image Processing, ICIP).
@inproceedings{33d35eeb662a49f89218165be4c45d4f,
title = "Pedestrian proposal generation using depth-aware scale estimation",
abstract = "In this work, we propose an efficient method that generates pedestrian proposals suitable for the autonomous vehicle. Our main intuition is that depth information provides an important cue to assign the scale of pedestrian proposals. Based on the observation that in a 3-D world coordinate the scales of pedestrians are almost similar, we formulate the scales of pedestrian patches by projecting 3-D models to an image plane with its corresponding depth. We also introduce a scale-aware binary description using both color and depth images. By using this descriptor, the regression models are trained to rank the pedestrian proposal candidates and adjust the proposal bounding boxes for an accurate localization. Our algorithm achieves significant performance gains compared to conventional proposal generation methods on the challenging KITTI dataset.",
author = "Kihong Park and Seungryong Kim and Kwanghoon Sohn",
year = "2018",
month = "2",
day = "20",
doi = "10.1109/ICIP.2017.8296641",
language = "English",
series = "Proceedings - International Conference on Image Processing, ICIP",
publisher = "IEEE Computer Society",
pages = "2045--2049",
booktitle = "2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings",
address = "United States",

}

Park, K, Kim, S & Sohn, K 2018, Pedestrian proposal generation using depth-aware scale estimation. in 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. Proceedings - International Conference on Image Processing, ICIP, vol. 2017-September, IEEE Computer Society, pp. 2045-2049, 24th IEEE International Conference on Image Processing, ICIP 2017, Beijing, China, 17/9/17. https://doi.org/10.1109/ICIP.2017.8296641

Pedestrian proposal generation using depth-aware scale estimation. / Park, Kihong; Kim, Seungryong; Sohn, Kwanghoon.

2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE Computer Society, 2018. p. 2045-2049 (Proceedings - International Conference on Image Processing, ICIP; Vol. 2017-September).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Pedestrian proposal generation using depth-aware scale estimation

AU - Park, Kihong

AU - Kim, Seungryong

AU - Sohn, Kwanghoon

PY - 2018/2/20

Y1 - 2018/2/20

N2 - In this work, we propose an efficient method that generates pedestrian proposals suitable for the autonomous vehicle. Our main intuition is that depth information provides an important cue to assign the scale of pedestrian proposals. Based on the observation that in a 3-D world coordinate the scales of pedestrians are almost similar, we formulate the scales of pedestrian patches by projecting 3-D models to an image plane with its corresponding depth. We also introduce a scale-aware binary description using both color and depth images. By using this descriptor, the regression models are trained to rank the pedestrian proposal candidates and adjust the proposal bounding boxes for an accurate localization. Our algorithm achieves significant performance gains compared to conventional proposal generation methods on the challenging KITTI dataset.

AB - In this work, we propose an efficient method that generates pedestrian proposals suitable for the autonomous vehicle. Our main intuition is that depth information provides an important cue to assign the scale of pedestrian proposals. Based on the observation that in a 3-D world coordinate the scales of pedestrians are almost similar, we formulate the scales of pedestrian patches by projecting 3-D models to an image plane with its corresponding depth. We also introduce a scale-aware binary description using both color and depth images. By using this descriptor, the regression models are trained to rank the pedestrian proposal candidates and adjust the proposal bounding boxes for an accurate localization. Our algorithm achieves significant performance gains compared to conventional proposal generation methods on the challenging KITTI dataset.

UR - http://www.scopus.com/inward/record.url?scp=85045313157&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85045313157&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2017.8296641

DO - 10.1109/ICIP.2017.8296641

M3 - Conference contribution

AN - SCOPUS:85045313157

T3 - Proceedings - International Conference on Image Processing, ICIP

SP - 2045

EP - 2049

BT - 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings

PB - IEEE Computer Society

ER -

Park K, Kim S, Sohn K. Pedestrian proposal generation using depth-aware scale estimation. In 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE Computer Society. 2018. p. 2045-2049. (Proceedings - International Conference on Image Processing, ICIP). https://doi.org/10.1109/ICIP.2017.8296641