TY  - JOUR
T1  - Fusion Approach for Content Based Image Retrieval by Utilizing Content and 
  Model Annotations
AU - Ambika, P. AU - Abdul Samath, J. 
JO  - Asian Journal of Information Technology
VL  - 12
IS  - 10
SP  - 325
EP  - 328
PY  - 2013
DA  - 2001/08/19
SN  - 1682-3915
DO  - ajit.2013.325.328
UR  - https://makhillpublications.co/view-article.php?doi=ajit.2013.325.328
KW  - Neural network
KW  -Genetic algorithm
KW  -population
KW  -crossover
KW  -mutation
AB  - Photography and television are playing a major role in facilitating the capture and communication of image data. Computers are the real engines of the imaging revolution bringing a range of techniques for digital image capture, processing, storage and transmission. In disparity to the text-based approach of such systems, CBIR maneuvers on a totally diverse attitude, salvaging stored images from a collection by matching features automatically mined from the images. Portrayal of content as well as its semantics is important in content based image retrieval. Even though the latest form of the CBIR System incorporates additional competent indexing techniques an improved user interface, the accuracy of retrieval is still low. Plentiful CBIR applications still depend on human categorized keywords for retrieval and deteriorating to emphasis on human interest point. This study proposes a fusion approach using neural network and genetic algorithm which exploits content and model annotations in to query refinement process. In order to reduce the semantic gap between the retrieved results and the user interest point, this study advises a two step approach which effectively incorporates the human interest point. Experimental results indicate that the fused approach doubled the accuracy and produced adequate results.
ER  -