Fast Query Point Movement Techniques with Efficient Feature Extraction for CBIR Systems

Sudhakar Murugesan, Edward Danso Ansong, Nivash Thirunavukarasu

Abstract


The explosive growth of image database has made the database management extremely tedious. Thus our approach focus on an efficient content based image retrieval(CBIR) technique that automatically search a particular image from a huge database with less iterations even in average and worst case. We propose a method with new index structure and query processing technique to improve the retrieval effectiveness and efficiency. This approach uses a combination of multi-resolution color, texture, and shape features to index the images. We have also considered strategies to minimize the effects of user’s inaccurate relevance feedback. Extensive experiments were simulated in realistic environments and the approach significantly reduced the number of required iterations and improved overall retrieval performance. The proposed system has used Corel database of 1000 images and the feature vectors of the query image were compared with feature vectors of the database and similar images were obtained.


 

Keywords: Fast Query Point, Movement Techniques, Efficient Feature, CBIR Systems


Full Text:

PDF

Refbacks

  • There are currently no refbacks.


ISSN 2408-770X (Print Version)

ISSN 2408-6231 (Online Version)