Sampling Operator to Learn the Scalable Correlation Filter for Visual Tracking

Minkyu Lee, Taeoh Kim, Yuseok Ban, Eungyeol Song, Sangyoun Lee

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

The correlation filter is suitable for tracking on account of its low computational complexity and promising performance. However, the number of available training samples is limited to the filter size, and the lack of samples leads to poor generalization. Moreover, spectral leakage degrades the filter quality. In this paper, we, therefore, propose a sampling operator for learning a scalable correlation filter in an enlarged window, whose size is larger than the object size. The scalable filter encodes the sparse frequency representation to reconstruct a larger filter with zeros outside of the object in the spatial domain. The sampling operator, which is composed of windowing and sampling operations, enables learning the scalable filter from a large window, and it suppresses spectral leakage. Our method was evaluated on the OTB-100, TC-128, and UAV-123 datasets and achieved the promising results in terms of precision and success rates.

Original languageEnglish
Article number8613772
Pages (from-to)11554-11564
Number of pages11
JournalIEEE Access
Volume7
DOIs
Publication statusPublished - 2019 Jan 1

Fingerprint

Mathematical operators
Sampling
Unmanned aerial vehicles (UAV)
Computational complexity

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Materials Science(all)
  • Engineering(all)

Cite this

Lee, Minkyu ; Kim, Taeoh ; Ban, Yuseok ; Song, Eungyeol ; Lee, Sangyoun. / Sampling Operator to Learn the Scalable Correlation Filter for Visual Tracking. In: IEEE Access. 2019 ; Vol. 7. pp. 11554-11564.
@article{1c1886359c7341ce9e7a1d3e6f66fcb0,
title = "Sampling Operator to Learn the Scalable Correlation Filter for Visual Tracking",
abstract = "The correlation filter is suitable for tracking on account of its low computational complexity and promising performance. However, the number of available training samples is limited to the filter size, and the lack of samples leads to poor generalization. Moreover, spectral leakage degrades the filter quality. In this paper, we, therefore, propose a sampling operator for learning a scalable correlation filter in an enlarged window, whose size is larger than the object size. The scalable filter encodes the sparse frequency representation to reconstruct a larger filter with zeros outside of the object in the spatial domain. The sampling operator, which is composed of windowing and sampling operations, enables learning the scalable filter from a large window, and it suppresses spectral leakage. Our method was evaluated on the OTB-100, TC-128, and UAV-123 datasets and achieved the promising results in terms of precision and success rates.",
author = "Minkyu Lee and Taeoh Kim and Yuseok Ban and Eungyeol Song and Sangyoun Lee",
year = "2019",
month = "1",
day = "1",
doi = "10.1109/ACCESS.2019.2892429",
language = "English",
volume = "7",
pages = "11554--11564",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Sampling Operator to Learn the Scalable Correlation Filter for Visual Tracking. / Lee, Minkyu; Kim, Taeoh; Ban, Yuseok; Song, Eungyeol; Lee, Sangyoun.

In: IEEE Access, Vol. 7, 8613772, 01.01.2019, p. 11554-11564.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Sampling Operator to Learn the Scalable Correlation Filter for Visual Tracking

AU - Lee, Minkyu

AU - Kim, Taeoh

AU - Ban, Yuseok

AU - Song, Eungyeol

AU - Lee, Sangyoun

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The correlation filter is suitable for tracking on account of its low computational complexity and promising performance. However, the number of available training samples is limited to the filter size, and the lack of samples leads to poor generalization. Moreover, spectral leakage degrades the filter quality. In this paper, we, therefore, propose a sampling operator for learning a scalable correlation filter in an enlarged window, whose size is larger than the object size. The scalable filter encodes the sparse frequency representation to reconstruct a larger filter with zeros outside of the object in the spatial domain. The sampling operator, which is composed of windowing and sampling operations, enables learning the scalable filter from a large window, and it suppresses spectral leakage. Our method was evaluated on the OTB-100, TC-128, and UAV-123 datasets and achieved the promising results in terms of precision and success rates.

AB - The correlation filter is suitable for tracking on account of its low computational complexity and promising performance. However, the number of available training samples is limited to the filter size, and the lack of samples leads to poor generalization. Moreover, spectral leakage degrades the filter quality. In this paper, we, therefore, propose a sampling operator for learning a scalable correlation filter in an enlarged window, whose size is larger than the object size. The scalable filter encodes the sparse frequency representation to reconstruct a larger filter with zeros outside of the object in the spatial domain. The sampling operator, which is composed of windowing and sampling operations, enables learning the scalable filter from a large window, and it suppresses spectral leakage. Our method was evaluated on the OTB-100, TC-128, and UAV-123 datasets and achieved the promising results in terms of precision and success rates.

UR - http://www.scopus.com/inward/record.url?scp=85061198812&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85061198812&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2019.2892429

DO - 10.1109/ACCESS.2019.2892429

M3 - Article

AN - SCOPUS:85061198812

VL - 7

SP - 11554

EP - 11564

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

M1 - 8613772

ER -