Accelerating arrays of linear classifiers using approximate range queries

Victor Lu, Ian Endres, Matei Stroila, John C. Hart

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

Abstract

Modern object detection methods apply binary linear classifiers on Euclidean feature vectors. This paper shows that projecting feature vectors onto a hypersphere allows an approximate range query to accelerate these detectors within acceptable levels of accuracy. The expense of constructing the k-d tree used by these range queries is justified when many detectors are used. We demonstrate our acceleration technique on several existing detection systems, including a state of the art logo detector, and show that approximate range queries can detect logos at least half as well at 11× the speed of the full fidelity method.

Original languageEnglish (US)
Title of host publication2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014
PublisherIEEE Computer Society
Pages255-260
Number of pages6
ISBN (Print)9781479949854
DOIs
StatePublished - Jan 1 2014
Event2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014 - Steamboat Springs, CO, United States
Duration: Mar 24 2014Mar 26 2014

Publication series

Name2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014

Other

Other2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014
CountryUnited States
CitySteamboat Springs, CO
Period3/24/143/26/14

Fingerprint

Classifiers
Detectors
Object detection

ASJC Scopus subject areas

  • Computer Science Applications
  • Computer Vision and Pattern Recognition

Cite this

Lu, V., Endres, I., Stroila, M., & Hart, J. C. (2014). Accelerating arrays of linear classifiers using approximate range queries. In 2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014 (pp. 255-260). [6836092] (2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014). IEEE Computer Society. https://doi.org/10.1109/WACV.2014.6836092

Accelerating arrays of linear classifiers using approximate range queries. / Lu, Victor; Endres, Ian; Stroila, Matei; Hart, John C.

2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014. IEEE Computer Society, 2014. p. 255-260 6836092 (2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014).

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

Lu, V, Endres, I, Stroila, M & Hart, JC 2014, Accelerating arrays of linear classifiers using approximate range queries. in 2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014., 6836092, 2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014, IEEE Computer Society, pp. 255-260, 2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014, Steamboat Springs, CO, United States, 3/24/14. https://doi.org/10.1109/WACV.2014.6836092
Lu V, Endres I, Stroila M, Hart JC. Accelerating arrays of linear classifiers using approximate range queries. In 2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014. IEEE Computer Society. 2014. p. 255-260. 6836092. (2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014). https://doi.org/10.1109/WACV.2014.6836092
Lu, Victor ; Endres, Ian ; Stroila, Matei ; Hart, John C. / Accelerating arrays of linear classifiers using approximate range queries. 2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014. IEEE Computer Society, 2014. pp. 255-260 (2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014).
@inproceedings{73fa1ffc25ee48939ef8aab4b15c3545,
title = "Accelerating arrays of linear classifiers using approximate range queries",
abstract = "Modern object detection methods apply binary linear classifiers on Euclidean feature vectors. This paper shows that projecting feature vectors onto a hypersphere allows an approximate range query to accelerate these detectors within acceptable levels of accuracy. The expense of constructing the k-d tree used by these range queries is justified when many detectors are used. We demonstrate our acceleration technique on several existing detection systems, including a state of the art logo detector, and show that approximate range queries can detect logos at least half as well at 11× the speed of the full fidelity method.",
author = "Victor Lu and Ian Endres and Matei Stroila and Hart, {John C.}",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/WACV.2014.6836092",
language = "English (US)",
isbn = "9781479949854",
series = "2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014",
publisher = "IEEE Computer Society",
pages = "255--260",
booktitle = "2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014",

}

TY - GEN

T1 - Accelerating arrays of linear classifiers using approximate range queries

AU - Lu, Victor

AU - Endres, Ian

AU - Stroila, Matei

AU - Hart, John C.

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Modern object detection methods apply binary linear classifiers on Euclidean feature vectors. This paper shows that projecting feature vectors onto a hypersphere allows an approximate range query to accelerate these detectors within acceptable levels of accuracy. The expense of constructing the k-d tree used by these range queries is justified when many detectors are used. We demonstrate our acceleration technique on several existing detection systems, including a state of the art logo detector, and show that approximate range queries can detect logos at least half as well at 11× the speed of the full fidelity method.

AB - Modern object detection methods apply binary linear classifiers on Euclidean feature vectors. This paper shows that projecting feature vectors onto a hypersphere allows an approximate range query to accelerate these detectors within acceptable levels of accuracy. The expense of constructing the k-d tree used by these range queries is justified when many detectors are used. We demonstrate our acceleration technique on several existing detection systems, including a state of the art logo detector, and show that approximate range queries can detect logos at least half as well at 11× the speed of the full fidelity method.

UR - http://www.scopus.com/inward/record.url?scp=84904615088&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84904615088&partnerID=8YFLogxK

U2 - 10.1109/WACV.2014.6836092

DO - 10.1109/WACV.2014.6836092

M3 - Conference contribution

AN - SCOPUS:84904615088

SN - 9781479949854

T3 - 2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014

SP - 255

EP - 260

BT - 2014 IEEE Winter Conference on Applications of Computer Vision, WACV 2014

PB - IEEE Computer Society

ER -