Utilization of Hierarchical and flat clustering in Content Based Image Retrieval

Priyanka Gupta


In this paper we are going to describe an image retrieval system whose input query is images and as output retrieves the images related to the image content similar to the query ignoring the similarities among the database images. Content based image retrieval means search will analyze the actual content of image rather than metadata for example attributes as tags, keywords and descriptions associated with the image. Here Actual content of the image may be considered as colors, textures, shapes and other information that can be derived from image itself. Main unique objective of this system is the utilization of hierarchical and k-means clustering techniques. Here proposed techniques consisting two stages-First in hierarchical clustering we filter most of the images in an unstructured large database and then applying clustered images to K-means such that system can return more better accurate results

Keywords: CBIR, hierarchical clustering, K-means, similarity

Full Text:


DOI: https://doi.org/10.26483/ijarcs.v3i7.1425


  • There are currently no refbacks.

Copyright (c) 2016 International Journal of Advanced Research in Computer Science