In recent years, covariance descriptors have received considerable attention as a strong representation of a set of points. In this research, we propose a new metric learning algorithm for covariance descriptors based on the Dykstra algorithm, in which the current solution is projected onto a half-space at each iteration, and which runs in O(n3) time. We empirically demonstrate that randomizing the order of half-spaces in the proposed Dykstra-based algorithm significantly accelerates convergence to the optimal solution. Furthermore, we show that the proposed approach yields promising experimental results for pattern recognition tasks.
Tomoki MATSUZAWA
Gunma University
Eisuke ITO
Gunma University
Raissa RELATOR
AIST Artificial Intelligence Research Center
Jun SESE
AIST Artificial Intelligence Research Center
Tsuyoshi KATO
Gunma University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Tomoki MATSUZAWA, Eisuke ITO, Raissa RELATOR, Jun SESE, Tsuyoshi KATO, "Stochastic Dykstra Algorithms for Distance Metric Learning with Covariance Descriptors" in IEICE TRANSACTIONS on Information,
vol. E100-D, no. 4, pp. 849-856, April 2017, doi: 10.1587/transinf.2016EDP7320.
Abstract: In recent years, covariance descriptors have received considerable attention as a strong representation of a set of points. In this research, we propose a new metric learning algorithm for covariance descriptors based on the Dykstra algorithm, in which the current solution is projected onto a half-space at each iteration, and which runs in O(n3) time. We empirically demonstrate that randomizing the order of half-spaces in the proposed Dykstra-based algorithm significantly accelerates convergence to the optimal solution. Furthermore, we show that the proposed approach yields promising experimental results for pattern recognition tasks.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2016EDP7320/_p
Copy
@ARTICLE{e100-d_4_849,
author={Tomoki MATSUZAWA, Eisuke ITO, Raissa RELATOR, Jun SESE, Tsuyoshi KATO, },
journal={IEICE TRANSACTIONS on Information},
title={Stochastic Dykstra Algorithms for Distance Metric Learning with Covariance Descriptors},
year={2017},
volume={E100-D},
number={4},
pages={849-856},
abstract={In recent years, covariance descriptors have received considerable attention as a strong representation of a set of points. In this research, we propose a new metric learning algorithm for covariance descriptors based on the Dykstra algorithm, in which the current solution is projected onto a half-space at each iteration, and which runs in O(n3) time. We empirically demonstrate that randomizing the order of half-spaces in the proposed Dykstra-based algorithm significantly accelerates convergence to the optimal solution. Furthermore, we show that the proposed approach yields promising experimental results for pattern recognition tasks.},
keywords={},
doi={10.1587/transinf.2016EDP7320},
ISSN={1745-1361},
month={April},}
Copy
TY - JOUR
TI - Stochastic Dykstra Algorithms for Distance Metric Learning with Covariance Descriptors
T2 - IEICE TRANSACTIONS on Information
SP - 849
EP - 856
AU - Tomoki MATSUZAWA
AU - Eisuke ITO
AU - Raissa RELATOR
AU - Jun SESE
AU - Tsuyoshi KATO
PY - 2017
DO - 10.1587/transinf.2016EDP7320
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 4
JA - IEICE TRANSACTIONS on Information
Y1 - April 2017
AB - In recent years, covariance descriptors have received considerable attention as a strong representation of a set of points. In this research, we propose a new metric learning algorithm for covariance descriptors based on the Dykstra algorithm, in which the current solution is projected onto a half-space at each iteration, and which runs in O(n3) time. We empirically demonstrate that randomizing the order of half-spaces in the proposed Dykstra-based algorithm significantly accelerates convergence to the optimal solution. Furthermore, we show that the proposed approach yields promising experimental results for pattern recognition tasks.
ER -