Personalised Image Search on Web Repository Using Partitioning Clustering Algorithm

Author(s):

Vighnesh.S , KCG College of Technology; Yogeshraj.M, KCG College of Technology; Sibi.S, KCG College of Technology

Keywords:

Clustering, Query-expansion, Database, Re-ranking, Image, Re-Ranking

Abstract:

The popular search engines like Google, Yahoo, and Bing are used to search and find different type of like text documents, PDF, images and videos. Usually image based searching process are carried out using the textual information associated with it and the final result has both relevant as well as irrelevant images from the user’s point of view. The expected output from these types of system is to get relevant result. In day-to-day life searching of text and images based on keyword and queries. Search using the keyword is considered to be easy and optimal methods for retrieving text, images and other details. The outcome from such a system is not always cent percent relevant. The reasons for irrelevant search results may be due to inappropriate query and user’s wrong perception. Image re-ranking, is an efficient way to lookup the results of web-based image search. Partitioning Clustering algorithm is used to remove noisy data. The retrieved images are ranked based on text similarity. When the user will select an query image from the collection, the left over images is re-ranked based on their shape, color and texture with the query image. The given query image in the database is maintained to produce the precise search results.

Other Details:

Manuscript Id	:	IJSTEV2I9059
Published in	:	Volume : 2, Issue : 9
Publication Date	:	01/04/2016
Page(s)	:	107-112

Download Article