Academic Journals Database
Disseminating quality controlled scientific knowledge

Bio-Medical Image Retrieval Using SVM

Author(s): S.Nithya, G.ShineLet

Journal: International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
ISSN 2278-1323

Volume: 1;
Issue: 10;
Start page: 014;
Date: 2012;
VIEW PDF   PDF DOWNLOAD PDF   Download PDF Original page

Keywords: Bio-Medical images | Low Level Features | Relevance Feedback (RF) | SupportVector Machine (SVM) Classifier.

This paper presents an image retrieval system in acomputer system for browsing, searching and retrieving imagesfrom a large database of digital images. The support vectormachine (SVM) is a concept based on learning methods thatevaluate data and identify patterns, usedfor classification and regression analysis. In this work, theprediction of database and query image depends on the outputsof a multiclass support vectormachine (SVM) classifier. Theirrelevant images are first to filter out, for similarity matching.The different low-level featuresmodalities are used in thisframework. When compare to the query image, the maindifficulty is to find unique feature. So an adaptive linearcombination of similarity matching and user’s feedbackinformation is used here. The predicted output of classifier maybe inexact in some cases. Hence, the feature weights arecalculated by considering both precision and rank orderinformation based on users feedback. As a result, using SVMclassifier with the retrieval performance clearly shows theadvantageof searching images based on similarity fusion andfiltering in terms of competent and success.
Why do you need a reservation system?      Affiliate Program