Feature based Image retrieval based on clustering, classification techniques using low level image features

Abstract

People are able to take photos using hand held devices and there is a massive
increase in the volume of photos digitally stored. However, this tremendous increase in the
number of digitally captured and stored images necessitates the development of advanced
techniques capable of classifying and effectively retrieving relevant images when needed.
Thus, Image retrieval has been popular for several years. There are different system designs
for image retrieval system. Image Retrieval is a technique of automatic indexing and
retrieving of images from a large data base.
Proposed system uses concept of data mining and image processing . Proposed system
which uses a well-known clustering algorithms k-means, fuzzy rule based classification and a
database indexing structure to facilitate retrieving relevant images in an efficie nt and
effective way. Colour histogram moments for RGB components and for HSV components are
used for colour information extraction. Gabor filter is used for texture feature extraction.
Image is clustered by its texture information, it is first level clustering which is done using Kmean clustering and secondly image is classified using its colour information which is done
using Fuzzy rule based classification. Use of clustering and classification reduces the search
space. The selection of cluster based on texture and colour information is done by various
experiments on different images. Fuzzy rule based classification allowed one data point to
belong to more than one class so retrieved image share good similarity than normal
clustering