Hyper least squares and its applications

Prasanna Rangarajan, Kenichi Kanatani, Hirotaka Niitsuma, Yasuyuki Sugaya

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

2 Citations (Scopus)

Abstract

We present a new form of least squares (LS), called "hyperLS", for geometric problems that frequently appear in computer vision applications. Doing rigorous error analysis, we maximize the accuracy by introducing a normalization that eliminates statistical bias up to second order noise terms. Our method yields a solution comparable to maximum likelihood (ML) without iterations, even in large noise situations where ML computation fails.

Original languageEnglish
Title of host publicationProceedings - International Conference on Pattern Recognition
Pages5-8
Number of pages4
DOIs
Publication statusPublished - 2010
Event2010 20th International Conference on Pattern Recognition, ICPR 2010 - Istanbul, Turkey
Duration: Aug 23 2010Aug 26 2010

Other

Other2010 20th International Conference on Pattern Recognition, ICPR 2010
CountryTurkey
CityIstanbul
Period8/23/108/26/10

Fingerprint

Maximum likelihood
Error analysis
Computer vision

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Rangarajan, P., Kanatani, K., Niitsuma, H., & Sugaya, Y. (2010). Hyper least squares and its applications. In Proceedings - International Conference on Pattern Recognition (pp. 5-8). [5597662] https://doi.org/10.1109/ICPR.2010.10

Hyper least squares and its applications. / Rangarajan, Prasanna; Kanatani, Kenichi; Niitsuma, Hirotaka; Sugaya, Yasuyuki.

Proceedings - International Conference on Pattern Recognition. 2010. p. 5-8 5597662.

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

Rangarajan, P, Kanatani, K, Niitsuma, H & Sugaya, Y 2010, Hyper least squares and its applications. in Proceedings - International Conference on Pattern Recognition., 5597662, pp. 5-8, 2010 20th International Conference on Pattern Recognition, ICPR 2010, Istanbul, Turkey, 8/23/10. https://doi.org/10.1109/ICPR.2010.10
Rangarajan P, Kanatani K, Niitsuma H, Sugaya Y. Hyper least squares and its applications. In Proceedings - International Conference on Pattern Recognition. 2010. p. 5-8. 5597662 https://doi.org/10.1109/ICPR.2010.10
Rangarajan, Prasanna ; Kanatani, Kenichi ; Niitsuma, Hirotaka ; Sugaya, Yasuyuki. / Hyper least squares and its applications. Proceedings - International Conference on Pattern Recognition. 2010. pp. 5-8
@inproceedings{bd8dc2cc23ae4b6bb0a9e5cd4121adf4,
title = "Hyper least squares and its applications",
abstract = "We present a new form of least squares (LS), called {"}hyperLS{"}, for geometric problems that frequently appear in computer vision applications. Doing rigorous error analysis, we maximize the accuracy by introducing a normalization that eliminates statistical bias up to second order noise terms. Our method yields a solution comparable to maximum likelihood (ML) without iterations, even in large noise situations where ML computation fails.",
author = "Prasanna Rangarajan and Kenichi Kanatani and Hirotaka Niitsuma and Yasuyuki Sugaya",
year = "2010",
doi = "10.1109/ICPR.2010.10",
language = "English",
isbn = "9780769541099",
pages = "5--8",
booktitle = "Proceedings - International Conference on Pattern Recognition",

}

TY - GEN

T1 - Hyper least squares and its applications

AU - Rangarajan, Prasanna

AU - Kanatani, Kenichi

AU - Niitsuma, Hirotaka

AU - Sugaya, Yasuyuki

PY - 2010

Y1 - 2010

N2 - We present a new form of least squares (LS), called "hyperLS", for geometric problems that frequently appear in computer vision applications. Doing rigorous error analysis, we maximize the accuracy by introducing a normalization that eliminates statistical bias up to second order noise terms. Our method yields a solution comparable to maximum likelihood (ML) without iterations, even in large noise situations where ML computation fails.

AB - We present a new form of least squares (LS), called "hyperLS", for geometric problems that frequently appear in computer vision applications. Doing rigorous error analysis, we maximize the accuracy by introducing a normalization that eliminates statistical bias up to second order noise terms. Our method yields a solution comparable to maximum likelihood (ML) without iterations, even in large noise situations where ML computation fails.

UR - http://www.scopus.com/inward/record.url?scp=78149471943&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78149471943&partnerID=8YFLogxK

U2 - 10.1109/ICPR.2010.10

DO - 10.1109/ICPR.2010.10

M3 - Conference contribution

AN - SCOPUS:78149471943

SN - 9780769541099

SP - 5

EP - 8

BT - Proceedings - International Conference on Pattern Recognition

ER -