An Image Reranking Model Based on Attributes and Visual Features Eliminating Duplication
Madhuri Mhaske1, Sachin Patil2
1Ms. Madhuri Mhaske, PG Scholar, Department of Computer Engineering, G. H. Raisoni College of Engineering and Management, Savitribai Phule Pune University, Chas, Ahmednagar (M.H), India.
2Prof. Sachin Patil, Assistant Professor, G.H. Raisoni College of Engineering and Management, Vagholi, Pune (M.H), India.
Manuscript received on 13 August 2016 | Revised Manuscript received on 20 August 2016 | Manuscript Published on 30 August 2016 | PP: 46-49 | Volume-5 Issue-6, August 2016 | Retrieval Number: F4654085616/16©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: An image search on internet is increasing day by day. Users type keywords in various search engines like Google, Yahoo, Bing etc for retrieval of relevant images. These search engines search the images from large pool of database. But as the keywords entered by user are generally short and ambiguous, different kinds of images are retrieved and sometimes these results are irrelevant. In this paper, semantic approach is proposed to solve this ambiguity. An image search reranking is definitely a superior approach over the text based image search. Using single modality for image searching is not sufficient as the different images have different features. This paper considers both the textual features as well as visual features for reranking. Attributes of images are classified into the groups. Based on those attributes from classifiers and the visual features of the images, each image is represented. The ranking score is used to evaluate the relevance of the image with query image. Hypergraph models these images based on the ranking scores .Content based image retrieval (CBIR) technique is used for extracting visual features. CBIR focuses on the content of the images such as color, texture, shape or any other information related with the images. Duplicate images found in search results are detected and eliminated by using SURF (Speeded Up Robust Feature) technique.
Keywords: Attribute, Hypergraph, CBIR, SURF. Etc
Scope of the Article: Image analysis and Processing