Authors | Hassan Farsi,Sajad Mohamadzadeh |
---|---|
Journal | Multimedia Tools and Applications |
Page number | 20895-20912 |
Serial number | 78 |
Volume number | 15 |
IF | 1.346 |
Paper Type | Full Paper |
Published At | 2019 |
Journal Grade | ISI |
Journal Type | Typographic |
Journal Country | Iran, Islamic Republic Of |
Journal Index | ISI،JCR،Scopus |
Abstract
As stored data and images on memory disks increase, image retrieval has a necessary task onimage processing. Although lots of researches have been reported for this task so far, semanticgap between low level features of images and human concept is still an important challenge oncontent-based image retrieval. For this task, a robust method is proposed by a combination ofconvolutional neural network and sparse representation, in which deep features are extractedby using CNN and sparse representation to increase retrieval speed and accuracy. Theproposed method has been tested on three common databases on image retrieval, namedCorel, ALOI and MPEG7. By computing metrics such as P(0.5), P(1) and ANMRR, exper-imental results show that the proposed method has achieved higher accuracy and better speedcompared to state-of-the-art methods.
tags: Content-based image retrieval.Deep learning.Convolutional neural networks.Sparse representation