Image Retrieval Framework Based on Dual Representation Descriptor

Yuichi YOSHIDA; Tsuyoshi TOYOFUKU

doi:10.1587/transinf.2017EDP7050

Image Retrieval Framework Based on Dual Representation Descriptor

Yuichi YOSHIDA, Tsuyoshi TOYOFUKU

Full Text Views

0

Share
Cite this

Summary :

Descriptor aggregation techniques such as the Fisher vector and vector of locally aggregated descriptors (VLAD) are used in most image retrieval frameworks. It takes some time to extract local descriptors, and the geometric verification requires storage if a real-valued descriptor such as SIFT is used. Moreover, if we apply binary descriptors to such a framework, the performance of image retrieval is not better than if we use a real-valued descriptor. Our approach tackles these issues by using a dual representation descriptor that has advantages of being both a real-valued and a binary descriptor. The real value of the dual representation descriptor is aggregated into a VLAD in order to achieve high accuracy in the image retrieval, and the binary one is used to find correspondences in the geometric verification stage in order to reduce the amount of storage needed. We implemented a dual representation descriptor extracted in semi-real time by using the CARD descriptor. We evaluated the accuracy of our image retrieval framework including the geometric verification on three datasets (holidays, ukbench and Stanford mobile visual search). The results indicate that our framework is as accurate as the framework that uses SIFT. In addition, the experiments show that the image retrieval speed and storage requirements of our framework are as efficient as those of a framework that uses ORB.

Publication: IEICE TRANSACTIONS on Information Vol.E100-D No.10 pp.2605-2613

Publication Date: 2017/10/01

Publicized: 2017/07/06

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2017EDP7050

Type of Manuscript: PAPER

Category: Image Processing and Video Processing

Authors

Yuichi YOSHIDA
DENSO IT Laboratory, Inc.
Tsuyoshi TOYOFUKU
DENSO IT Laboratory, Inc.

Keyword

image retrieval, local descriptor

Cite this

Copy

Yuichi YOSHIDA, Tsuyoshi TOYOFUKU, "Image Retrieval Framework Based on Dual Representation Descriptor" in IEICE TRANSACTIONS on Information, vol. E100-D, no. 10, pp. 2605-2613, October 2017, doi: 10.1587/transinf.2017EDP7050.
Abstract: Descriptor aggregation techniques such as the Fisher vector and vector of locally aggregated descriptors (VLAD) are used in most image retrieval frameworks. It takes some time to extract local descriptors, and the geometric verification requires storage if a real-valued descriptor such as SIFT is used. Moreover, if we apply binary descriptors to such a framework, the performance of image retrieval is not better than if we use a real-valued descriptor. Our approach tackles these issues by using a dual representation descriptor that has advantages of being both a real-valued and a binary descriptor. The real value of the dual representation descriptor is aggregated into a VLAD in order to achieve high accuracy in the image retrieval, and the binary one is used to find correspondences in the geometric verification stage in order to reduce the amount of storage needed. We implemented a dual representation descriptor extracted in semi-real time by using the CARD descriptor. We evaluated the accuracy of our image retrieval framework including the geometric verification on three datasets (holidays, ukbench and Stanford mobile visual search). The results indicate that our framework is as accurate as the framework that uses SIFT. In addition, the experiments show that the image retrieval speed and storage requirements of our framework are as efficient as those of a framework that uses ORB.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.2017EDP7050/_p

Copy

@ARTICLE{e100-d_10_2605,
author={Yuichi YOSHIDA, Tsuyoshi TOYOFUKU, },
journal={IEICE TRANSACTIONS on Information},
title={Image Retrieval Framework Based on Dual Representation Descriptor},
year={2017},
volume={E100-D},
number={10},
pages={2605-2613},
abstract={Descriptor aggregation techniques such as the Fisher vector and vector of locally aggregated descriptors (VLAD) are used in most image retrieval frameworks. It takes some time to extract local descriptors, and the geometric verification requires storage if a real-valued descriptor such as SIFT is used. Moreover, if we apply binary descriptors to such a framework, the performance of image retrieval is not better than if we use a real-valued descriptor. Our approach tackles these issues by using a dual representation descriptor that has advantages of being both a real-valued and a binary descriptor. The real value of the dual representation descriptor is aggregated into a VLAD in order to achieve high accuracy in the image retrieval, and the binary one is used to find correspondences in the geometric verification stage in order to reduce the amount of storage needed. We implemented a dual representation descriptor extracted in semi-real time by using the CARD descriptor. We evaluated the accuracy of our image retrieval framework including the geometric verification on three datasets (holidays, ukbench and Stanford mobile visual search). The results indicate that our framework is as accurate as the framework that uses SIFT. In addition, the experiments show that the image retrieval speed and storage requirements of our framework are as efficient as those of a framework that uses ORB.},
keywords={},
doi={10.1587/transinf.2017EDP7050},
ISSN={1745-1361},
month={October},}

Copy

TY - JOUR
TI - Image Retrieval Framework Based on Dual Representation Descriptor
T2 - IEICE TRANSACTIONS on Information
SP - 2605
EP - 2613
AU - Yuichi YOSHIDA
AU - Tsuyoshi TOYOFUKU
PY - 2017
DO - 10.1587/transinf.2017EDP7050
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2017
AB - Descriptor aggregation techniques such as the Fisher vector and vector of locally aggregated descriptors (VLAD) are used in most image retrieval frameworks. It takes some time to extract local descriptors, and the geometric verification requires storage if a real-valued descriptor such as SIFT is used. Moreover, if we apply binary descriptors to such a framework, the performance of image retrieval is not better than if we use a real-valued descriptor. Our approach tackles these issues by using a dual representation descriptor that has advantages of being both a real-valued and a binary descriptor. The real value of the dual representation descriptor is aggregated into a VLAD in order to achieve high accuracy in the image retrieval, and the binary one is used to find correspondences in the geometric verification stage in order to reduce the amount of storage needed. We implemented a dual representation descriptor extracted in semi-real time by using the CARD descriptor. We evaluated the accuracy of our image retrieval framework including the geometric verification on three datasets (holidays, ukbench and Stanford mobile visual search). The results indicate that our framework is as accurate as the framework that uses SIFT. In addition, the experiments show that the image retrieval speed and storage requirements of our framework are as efficient as those of a framework that uses ORB.
ER -