INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
A NOVEL APPROACH FOR DISTANCE METRIC LEARNING for MULTI-MODAL IMAGE RETRIEVAL
Authors Name:
M.Vinuthna
, CH. Ratna Jyothi , A. Yuva Krishna
Unique Id:
IJSDR1709027
Published In:
Volume 2 Issue 9, September-2017
Abstract:
One of the core research problems in multimedia retrieval is to seek an effective distance metric/function for computing similarity of two objects in content-based multimedia retrieval tasks. Distance metric learning (DML) is an important technique to improve similarity search in content-based image retrieval. The paper proposes the current online multi-modal distance metric learning (OMDML) with another component of expansion to illuminate the Image equivocalness issue utilizing fundamental expectation of proposing this model of framework is to comment on/label the images with some physically characterized ideas for learning a natural space, utilizing visual and logical features. All the most especially, by creating the framework to sustain the dormant vectors into existing classification portrayals, it can be authorize to be used of image comment, which is considered as the required issue in image recovery. As an expansion to the accessible model, we suggest and include the substance highlight of the issue of understanding the vagueness. Online multi-modal distance metric learning framework gives a superior results of substance based image recovery show. We investigate a completely unique theme of on-line multi-modal distance metric learning (OMDML), that explores a unified two-level on-line learning scheme: (i) it learns to optimize a distance metric on every individual feature space; and (ii) then it learns to And the optimal combination of diverse types of features. To further reduce the expensive cost of DML on high-dimensional feature space, we propose a low-rank OMDML algorithm which not only significantly reduces the computational cost but also retains highly competing or even better learning accuracy.
"A NOVEL APPROACH FOR DISTANCE METRIC LEARNING for MULTI-MODAL IMAGE RETRIEVAL", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.2, Issue 9, page no.159 - 163, September-2017, Available :http://www.ijsdr.org/papers/IJSDR1709027.pdf
Downloads:
000337362
Publication Details:
Published Paper ID: IJSDR1709027
Registration ID:170770
Published In: Volume 2 Issue 9, September-2017
DOI (Digital Object Identifier):
Page No: 159 - 163
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn