Robust visual tracking using oblique random forests

Le Zhang, Jagannadan Varadarajan, Ponnuthurai Nagaratnam Suganthan, Narendra Ahuja, Pierre Moulin

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

Abstract

Random forest has emerged as a powerful classification technique with promising results in various vision tasks including image classification, pose estimation and object detection. However, current techniques have shown little improvements in visual tracking as they mostly rely on piece wise orthogonal hyperplanes to create decision nodes and lack a robust incremental learning mechanism that is much needed for online tracking. In this paper, we propose a discriminative tracker based on a novel incremental oblique random forest. Unlike conventional orthogonal decision trees that use a single feature and heuristic measures to obtain a split at each node, we propose to use a more powerful proximal SVM to obtain oblique hyperplanes to capture the geometric structure of the data better. The resulting decision surface is not restricted to be axis aligned, and hence has the ability to represent and classify the input data better. Furthermore, in order to generalize to online tracking scenarios, we derive incremental update steps that enable the hyperplanes in each node to be updated recursively, efficiently and in a closed-form fashion. We demonstrate the effectiveness of our method using two large scale benchmark datasets (OTB-51 and OTB-100) and show that our method gives competitive results on several challenging cases by relying on simple HOG features as well as in combination with more sophisticated deep neural network based models. The implementations of the proposed random forest are available at https://github.com/ZhangLeUestc/Incremental-Oblique-Random-Forest.

Original languageEnglish (US)
Title of host publicationProceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages5825-5834
Number of pages10
ISBN (Electronic)9781538604571
DOIs
StatePublished - Nov 6 2017
Event30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017 - Honolulu, United States
Duration: Jul 21 2017Jul 26 2017

Publication series

NameProceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017
Volume2017-January

Other

Other30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017
CountryUnited States
CityHonolulu
Period7/21/177/26/17

Fingerprint

Image classification
Decision trees
Deep neural networks
Object detection

ASJC Scopus subject areas

  • Signal Processing
  • Computer Vision and Pattern Recognition

Cite this

Zhang, L., Varadarajan, J., Suganthan, P. N., Ahuja, N., & Moulin, P. (2017). Robust visual tracking using oblique random forests. In Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017 (pp. 5825-5834). (Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017; Vol. 2017-January). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CVPR.2017.617

Robust visual tracking using oblique random forests. / Zhang, Le; Varadarajan, Jagannadan; Suganthan, Ponnuthurai Nagaratnam; Ahuja, Narendra; Moulin, Pierre.

Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Institute of Electrical and Electronics Engineers Inc., 2017. p. 5825-5834 (Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017; Vol. 2017-January).

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

Zhang, L, Varadarajan, J, Suganthan, PN, Ahuja, N & Moulin, P 2017, Robust visual tracking using oblique random forests. in Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, vol. 2017-January, Institute of Electrical and Electronics Engineers Inc., pp. 5825-5834, 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, Honolulu, United States, 7/21/17. https://doi.org/10.1109/CVPR.2017.617
Zhang L, Varadarajan J, Suganthan PN, Ahuja N, Moulin P. Robust visual tracking using oblique random forests. In Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 5825-5834. (Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017). https://doi.org/10.1109/CVPR.2017.617
Zhang, Le ; Varadarajan, Jagannadan ; Suganthan, Ponnuthurai Nagaratnam ; Ahuja, Narendra ; Moulin, Pierre. / Robust visual tracking using oblique random forests. Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 5825-5834 (Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017).
@inproceedings{ae1a8f65d97e4f46b9a0263b506908c8,
title = "Robust visual tracking using oblique random forests",
abstract = "Random forest has emerged as a powerful classification technique with promising results in various vision tasks including image classification, pose estimation and object detection. However, current techniques have shown little improvements in visual tracking as they mostly rely on piece wise orthogonal hyperplanes to create decision nodes and lack a robust incremental learning mechanism that is much needed for online tracking. In this paper, we propose a discriminative tracker based on a novel incremental oblique random forest. Unlike conventional orthogonal decision trees that use a single feature and heuristic measures to obtain a split at each node, we propose to use a more powerful proximal SVM to obtain oblique hyperplanes to capture the geometric structure of the data better. The resulting decision surface is not restricted to be axis aligned, and hence has the ability to represent and classify the input data better. Furthermore, in order to generalize to online tracking scenarios, we derive incremental update steps that enable the hyperplanes in each node to be updated recursively, efficiently and in a closed-form fashion. We demonstrate the effectiveness of our method using two large scale benchmark datasets (OTB-51 and OTB-100) and show that our method gives competitive results on several challenging cases by relying on simple HOG features as well as in combination with more sophisticated deep neural network based models. The implementations of the proposed random forest are available at https://github.com/ZhangLeUestc/Incremental-Oblique-Random-Forest.",
author = "Le Zhang and Jagannadan Varadarajan and Suganthan, {Ponnuthurai Nagaratnam} and Narendra Ahuja and Pierre Moulin",
year = "2017",
month = "11",
day = "6",
doi = "10.1109/CVPR.2017.617",
language = "English (US)",
series = "Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "5825--5834",
booktitle = "Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017",
address = "United States",

}

TY - GEN

T1 - Robust visual tracking using oblique random forests

AU - Zhang, Le

AU - Varadarajan, Jagannadan

AU - Suganthan, Ponnuthurai Nagaratnam

AU - Ahuja, Narendra

AU - Moulin, Pierre

PY - 2017/11/6

Y1 - 2017/11/6

N2 - Random forest has emerged as a powerful classification technique with promising results in various vision tasks including image classification, pose estimation and object detection. However, current techniques have shown little improvements in visual tracking as they mostly rely on piece wise orthogonal hyperplanes to create decision nodes and lack a robust incremental learning mechanism that is much needed for online tracking. In this paper, we propose a discriminative tracker based on a novel incremental oblique random forest. Unlike conventional orthogonal decision trees that use a single feature and heuristic measures to obtain a split at each node, we propose to use a more powerful proximal SVM to obtain oblique hyperplanes to capture the geometric structure of the data better. The resulting decision surface is not restricted to be axis aligned, and hence has the ability to represent and classify the input data better. Furthermore, in order to generalize to online tracking scenarios, we derive incremental update steps that enable the hyperplanes in each node to be updated recursively, efficiently and in a closed-form fashion. We demonstrate the effectiveness of our method using two large scale benchmark datasets (OTB-51 and OTB-100) and show that our method gives competitive results on several challenging cases by relying on simple HOG features as well as in combination with more sophisticated deep neural network based models. The implementations of the proposed random forest are available at https://github.com/ZhangLeUestc/Incremental-Oblique-Random-Forest.

AB - Random forest has emerged as a powerful classification technique with promising results in various vision tasks including image classification, pose estimation and object detection. However, current techniques have shown little improvements in visual tracking as they mostly rely on piece wise orthogonal hyperplanes to create decision nodes and lack a robust incremental learning mechanism that is much needed for online tracking. In this paper, we propose a discriminative tracker based on a novel incremental oblique random forest. Unlike conventional orthogonal decision trees that use a single feature and heuristic measures to obtain a split at each node, we propose to use a more powerful proximal SVM to obtain oblique hyperplanes to capture the geometric structure of the data better. The resulting decision surface is not restricted to be axis aligned, and hence has the ability to represent and classify the input data better. Furthermore, in order to generalize to online tracking scenarios, we derive incremental update steps that enable the hyperplanes in each node to be updated recursively, efficiently and in a closed-form fashion. We demonstrate the effectiveness of our method using two large scale benchmark datasets (OTB-51 and OTB-100) and show that our method gives competitive results on several challenging cases by relying on simple HOG features as well as in combination with more sophisticated deep neural network based models. The implementations of the proposed random forest are available at https://github.com/ZhangLeUestc/Incremental-Oblique-Random-Forest.

UR - http://www.scopus.com/inward/record.url?scp=85044391980&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85044391980&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2017.617

DO - 10.1109/CVPR.2017.617

M3 - Conference contribution

AN - SCOPUS:85044391980

T3 - Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017

SP - 5825

EP - 5834

BT - Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -