Adaptive window strategy for high-speed and robust KLT feature tracker

Nirmala Ramakrishnan, Thambipillai Srikanthan, Siew Kei Lam, Gauri Ravindra Tulsulkar

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

3 Citations (Scopus)

Abstract

The Kanade-Lucas-Tomasi tracking (KLT) algorithm is widely used for local tracking of features. As it employs a translation model to find the feature tracks, KLT is not robust in the presence of distortions around the feature resulting in high inaccuracies in the tracks. In this paper we show that the window size in KLT must vary to adapt to the presence of distortions around each feature point in order to increase the number of useful tracks and minimize noisy ones. We propose an adaptive window size strategy for KLT that uses the KLT iterations as an indicator of the quality of the tracks to determine near-optimal window sizes, thereby significantly improving its robustness to distortions. Our evaluations with a well-known tracking dataset show that the proposed adaptive strategy outperforms the conventional fixed-window KLT in terms of robustness. In addition, compared to the well-known affine KLT, our method achieves comparable robustness at an average runtime speedup of 7x.

Original languageEnglish
Title of host publicationImage and Video Technology - 7th Pacific-Rim Symposium, PSIVT 2015, Revised Selected Papers
PublisherSpringer Verlag
Pages355-367
Number of pages13
ISBN (Print)9783319294506
DOIs
Publication statusPublished - 01-01-2016
Externally publishedYes
Event7th Pacific-Rim Symposium on Image and Video Technology, PSIVT 2015 - Auckland, New Zealand
Duration: 25-11-201527-11-2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9431
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference7th Pacific-Rim Symposium on Image and Video Technology, PSIVT 2015
CountryNew Zealand
CityAuckland
Period25-11-1527-11-15

Fingerprint

Feature Tracking
High Speed
Robustness
Strategy
Adaptive Strategies
Feature Point
Speedup
Vary
Minimise
Iteration

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Ramakrishnan, N., Srikanthan, T., Lam, S. K., & Tulsulkar, G. R. (2016). Adaptive window strategy for high-speed and robust KLT feature tracker. In Image and Video Technology - 7th Pacific-Rim Symposium, PSIVT 2015, Revised Selected Papers (pp. 355-367). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9431). Springer Verlag. https://doi.org/10.1007/978-3-319-29451-3_29
Ramakrishnan, Nirmala ; Srikanthan, Thambipillai ; Lam, Siew Kei ; Tulsulkar, Gauri Ravindra. / Adaptive window strategy for high-speed and robust KLT feature tracker. Image and Video Technology - 7th Pacific-Rim Symposium, PSIVT 2015, Revised Selected Papers. Springer Verlag, 2016. pp. 355-367 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{eb3edd3964e8437999cd70faa5d5db60,
title = "Adaptive window strategy for high-speed and robust KLT feature tracker",
abstract = "The Kanade-Lucas-Tomasi tracking (KLT) algorithm is widely used for local tracking of features. As it employs a translation model to find the feature tracks, KLT is not robust in the presence of distortions around the feature resulting in high inaccuracies in the tracks. In this paper we show that the window size in KLT must vary to adapt to the presence of distortions around each feature point in order to increase the number of useful tracks and minimize noisy ones. We propose an adaptive window size strategy for KLT that uses the KLT iterations as an indicator of the quality of the tracks to determine near-optimal window sizes, thereby significantly improving its robustness to distortions. Our evaluations with a well-known tracking dataset show that the proposed adaptive strategy outperforms the conventional fixed-window KLT in terms of robustness. In addition, compared to the well-known affine KLT, our method achieves comparable robustness at an average runtime speedup of 7x.",
author = "Nirmala Ramakrishnan and Thambipillai Srikanthan and Lam, {Siew Kei} and Tulsulkar, {Gauri Ravindra}",
year = "2016",
month = "1",
day = "1",
doi = "10.1007/978-3-319-29451-3_29",
language = "English",
isbn = "9783319294506",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "355--367",
booktitle = "Image and Video Technology - 7th Pacific-Rim Symposium, PSIVT 2015, Revised Selected Papers",
address = "Germany",

}

Ramakrishnan, N, Srikanthan, T, Lam, SK & Tulsulkar, GR 2016, Adaptive window strategy for high-speed and robust KLT feature tracker. in Image and Video Technology - 7th Pacific-Rim Symposium, PSIVT 2015, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9431, Springer Verlag, pp. 355-367, 7th Pacific-Rim Symposium on Image and Video Technology, PSIVT 2015, Auckland, New Zealand, 25-11-15. https://doi.org/10.1007/978-3-319-29451-3_29

Adaptive window strategy for high-speed and robust KLT feature tracker. / Ramakrishnan, Nirmala; Srikanthan, Thambipillai; Lam, Siew Kei; Tulsulkar, Gauri Ravindra.

Image and Video Technology - 7th Pacific-Rim Symposium, PSIVT 2015, Revised Selected Papers. Springer Verlag, 2016. p. 355-367 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9431).

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

TY - GEN

T1 - Adaptive window strategy for high-speed and robust KLT feature tracker

AU - Ramakrishnan, Nirmala

AU - Srikanthan, Thambipillai

AU - Lam, Siew Kei

AU - Tulsulkar, Gauri Ravindra

PY - 2016/1/1

Y1 - 2016/1/1

N2 - The Kanade-Lucas-Tomasi tracking (KLT) algorithm is widely used for local tracking of features. As it employs a translation model to find the feature tracks, KLT is not robust in the presence of distortions around the feature resulting in high inaccuracies in the tracks. In this paper we show that the window size in KLT must vary to adapt to the presence of distortions around each feature point in order to increase the number of useful tracks and minimize noisy ones. We propose an adaptive window size strategy for KLT that uses the KLT iterations as an indicator of the quality of the tracks to determine near-optimal window sizes, thereby significantly improving its robustness to distortions. Our evaluations with a well-known tracking dataset show that the proposed adaptive strategy outperforms the conventional fixed-window KLT in terms of robustness. In addition, compared to the well-known affine KLT, our method achieves comparable robustness at an average runtime speedup of 7x.

AB - The Kanade-Lucas-Tomasi tracking (KLT) algorithm is widely used for local tracking of features. As it employs a translation model to find the feature tracks, KLT is not robust in the presence of distortions around the feature resulting in high inaccuracies in the tracks. In this paper we show that the window size in KLT must vary to adapt to the presence of distortions around each feature point in order to increase the number of useful tracks and minimize noisy ones. We propose an adaptive window size strategy for KLT that uses the KLT iterations as an indicator of the quality of the tracks to determine near-optimal window sizes, thereby significantly improving its robustness to distortions. Our evaluations with a well-known tracking dataset show that the proposed adaptive strategy outperforms the conventional fixed-window KLT in terms of robustness. In addition, compared to the well-known affine KLT, our method achieves comparable robustness at an average runtime speedup of 7x.

UR - http://www.scopus.com/inward/record.url?scp=84958971831&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958971831&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-29451-3_29

DO - 10.1007/978-3-319-29451-3_29

M3 - Conference contribution

AN - SCOPUS:84958971831

SN - 9783319294506

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 355

EP - 367

BT - Image and Video Technology - 7th Pacific-Rim Symposium, PSIVT 2015, Revised Selected Papers

PB - Springer Verlag

ER -

Ramakrishnan N, Srikanthan T, Lam SK, Tulsulkar GR. Adaptive window strategy for high-speed and robust KLT feature tracker. In Image and Video Technology - 7th Pacific-Rim Symposium, PSIVT 2015, Revised Selected Papers. Springer Verlag. 2016. p. 355-367. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-29451-3_29