Aljuaidi, Reem and Su, Jing and Dahyot, Rozenn (2019) Mini-Batch VLAD for Visual Place Retrieval. 2019 30th Irish Signals and Systems Conference (ISSC). pp. 1-6.

Preview

Download (482kB) | Preview

Abstract

This study investigates the visual place retrieval of an image query using a geotagged image dataset. Vector of Locally Aggregated Descriptors (VLAD) is one of the local features that can be used for image place recognition. VLAD describes an image by the difference of its local feature descriptors from an already computed codebook. Generally, a visual codebook is generated from k-means clustering of the descriptors. However, the dimensionality of visual features is not trivial and the computational load of sample distances in a large image dataset is challenging. In order to design an accurate image retrieval method with affordable computation expenses, we propose to use the mini-batch k-means clustering to compute VLAD descriptor(MB-VLAD). The proposed MB�VLAD technique shows advantage in retrieval accuracy in comparison with the state of the art techniques.

Item Type:	Article
Keywords:	feature extraction; content-based image retrieval; image processing;
Academic Unit:	Faculty of Science and Engineering > Computer Science
Item ID:	15129
Identification Number:	https://doi.org/10.1109/ISSC.2019.8904931
Depositing User:	Rozenn Dahyot
Date Deposited:	14 Dec 2021 16:49
Journal or Publication Title:	2019 30th Irish Signals and Systems Conference (ISSC)
Publisher:	IEEE
Refereed:	Yes
URI:	Publisher
Use Licence:	This item is available under a Creative Commons Attribution Non Commercial Share Alike Licence (CC BY-NC-SA). Details of this licence are available here

Repository Staff Only(login required)

Item control page

Downloads

Origin of downloads