Improving Hough Based Pedestrian Detection Accuracy by Using Segmentation and Pose Subspaces

Jarich VANSTEENBERGE; Masayuki MUKUNOKI; Michihiko MINOH

doi:10.1587/transinf.2014EDP7092

Improving Hough Based Pedestrian Detection Accuracy by Using Segmentation and Pose Subspaces

Jarich VANSTEENBERGE, Masayuki MUKUNOKI, Michihiko MINOH

Full Text Views

0

Share
Cite this

Summary :

The Hough voting framework is a popular approach to parts based pedestrian detection. It works by allowing image features to vote for the positions and scales of pedestrians within a test image. Each vote is cast independently from other votes, which allows for strong occlusion robustness. However this approach can produce false pedestrian detections by accumulating votes inconsistent with each other, especially in cluttered scenes such as typical street scenes. This work aims to reduce the sensibility to clutter in the Hough voting framework. Our idea is to use object segmentation and object pose parameters to enforce votes' consistency both at training and testing time. Specifically, we use segmentation and pose parameters to guide the learning of a pedestrian model able to cast mutually consistent votes. At test time, each candidate detection's support votes are looked upon from a segmentation and pose viewpoints to measure their level of agreement. We show that this measure provides an efficient way to discriminate between true and false detections. We tested our method on four challenging pedestrian datasets. Our method shows clear improvements over the original Hough based detectors and performs on par with recent enhanced Hough based detectors. In addition, our method can perform segmentation and pose estimation as byproducts of the detection process.

Publication: IEICE TRANSACTIONS on Information Vol.E97-D No.10 pp.2760-2768

Publication Date: 2014/10/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2014EDP7092

Type of Manuscript: PAPER

Category: Image Recognition, Computer Vision

Authors

Jarich VANSTEENBERGE
  Graduate School of Informatics, Kyoto University
Masayuki MUKUNOKI
  Kyoto University
Michihiko MINOH
  Kyoto University

Keyword

Hough based detections, pedestrian segmentation, pose estimation, Random Forest, kPCA

Cite this

Copy

Jarich VANSTEENBERGE, Masayuki MUKUNOKI, Michihiko MINOH, "Improving Hough Based Pedestrian Detection Accuracy by Using Segmentation and Pose Subspaces" in IEICE TRANSACTIONS on Information, vol. E97-D, no. 10, pp. 2760-2768, October 2014, doi: 10.1587/transinf.2014EDP7092.
Abstract: The Hough voting framework is a popular approach to parts based pedestrian detection. It works by allowing image features to vote for the positions and scales of pedestrians within a test image. Each vote is cast independently from other votes, which allows for strong occlusion robustness. However this approach can produce false pedestrian detections by accumulating votes inconsistent with each other, especially in cluttered scenes such as typical street scenes. This work aims to reduce the sensibility to clutter in the Hough voting framework. Our idea is to use object segmentation and object pose parameters to enforce votes' consistency both at training and testing time. Specifically, we use segmentation and pose parameters to guide the learning of a pedestrian model able to cast mutually consistent votes. At test time, each candidate detection's support votes are looked upon from a segmentation and pose viewpoints to measure their level of agreement. We show that this measure provides an efficient way to discriminate between true and false detections. We tested our method on four challenging pedestrian datasets. Our method shows clear improvements over the original Hough based detectors and performs on par with recent enhanced Hough based detectors. In addition, our method can perform segmentation and pose estimation as byproducts of the detection process.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.2014EDP7092/_p

Copy

@ARTICLE{e97-d_10_2760,
author={Jarich VANSTEENBERGE, Masayuki MUKUNOKI, Michihiko MINOH, },
journal={IEICE TRANSACTIONS on Information},
title={Improving Hough Based Pedestrian Detection Accuracy by Using Segmentation and Pose Subspaces},
year={2014},
volume={E97-D},
number={10},
pages={2760-2768},
abstract={The Hough voting framework is a popular approach to parts based pedestrian detection. It works by allowing image features to vote for the positions and scales of pedestrians within a test image. Each vote is cast independently from other votes, which allows for strong occlusion robustness. However this approach can produce false pedestrian detections by accumulating votes inconsistent with each other, especially in cluttered scenes such as typical street scenes. This work aims to reduce the sensibility to clutter in the Hough voting framework. Our idea is to use object segmentation and object pose parameters to enforce votes' consistency both at training and testing time. Specifically, we use segmentation and pose parameters to guide the learning of a pedestrian model able to cast mutually consistent votes. At test time, each candidate detection's support votes are looked upon from a segmentation and pose viewpoints to measure their level of agreement. We show that this measure provides an efficient way to discriminate between true and false detections. We tested our method on four challenging pedestrian datasets. Our method shows clear improvements over the original Hough based detectors and performs on par with recent enhanced Hough based detectors. In addition, our method can perform segmentation and pose estimation as byproducts of the detection process.},
keywords={},
doi={10.1587/transinf.2014EDP7092},
ISSN={1745-1361},
month={October},}

Copy

TY - JOUR
TI - Improving Hough Based Pedestrian Detection Accuracy by Using Segmentation and Pose Subspaces
T2 - IEICE TRANSACTIONS on Information
SP - 2760
EP - 2768
AU - Jarich VANSTEENBERGE
AU - Masayuki MUKUNOKI
AU - Michihiko MINOH
PY - 2014
DO - 10.1587/transinf.2014EDP7092
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E97-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2014
AB - The Hough voting framework is a popular approach to parts based pedestrian detection. It works by allowing image features to vote for the positions and scales of pedestrians within a test image. Each vote is cast independently from other votes, which allows for strong occlusion robustness. However this approach can produce false pedestrian detections by accumulating votes inconsistent with each other, especially in cluttered scenes such as typical street scenes. This work aims to reduce the sensibility to clutter in the Hough voting framework. Our idea is to use object segmentation and object pose parameters to enforce votes' consistency both at training and testing time. Specifically, we use segmentation and pose parameters to guide the learning of a pedestrian model able to cast mutually consistent votes. At test time, each candidate detection's support votes are looked upon from a segmentation and pose viewpoints to measure their level of agreement. We show that this measure provides an efficient way to discriminate between true and false detections. We tested our method on four challenging pedestrian datasets. Our method shows clear improvements over the original Hough based detectors and performs on par with recent enhanced Hough based detectors. In addition, our method can perform segmentation and pose estimation as byproducts of the detection process.
ER -