文章基本信息

标题：Improve the Effectiveness of Image Retrieval by Combining the Optimal Distance and Linear Discriminant Analysis
本地全文：下载
作者：Phuong Nguyen Thi Lan ; Tao Ngo Quoc ; Quynh Dao Thi Thuy 等
期刊名称：International Journal of Advanced Computer Science and Applications(IJACSA)
印刷版ISSN：2158-107X
电子版ISSN：2156-5570
出版年度：2021
卷号：12
期号：2
页码：46-52
DOI：10.14569/IJACSA.2021.0120206
出版社：Science and Information Society (SAI)
摘要：In image retrieval with relevant feedback, classification and distance calculation have a great influence on image retrieval accuracy. In this paper, we propose an image retrieval method, called ODLDA (Image Retrieval using the optimal distance and linear discriminant analysis). The proposed method can effectively exploit user’s feedback from relevant and irrelevant image sets, which uses linear discriminant analysis to find a linear projection with an improved similarity measure. The experimental results performed on the two benchmark datasets have confirmed the superiority of the proposed method.
关键词：Content-based image retrieval; deep learning; similarity measures; Mahalanobis metric distance; linear discriminant analysis