A novel knowledge management scheme for IRS

This paper presents a brief overview of knowledge exploitation in image retrieval system (IRS). We also proposed framework for knowledge based image retrieval system which integrates image mining, hybrid knowledge representation techniques and knowledge log. This system derived knowledge from image database by image mining, which is represented by using hybrid methodology of semantic net and frame. We introduce knowledge log which maintain previous query result and conclude conclusions from them for future use by using learning system associated with it. We work on small image database. Experimental result shows improved system performance which is studied with respect to parameters like precision, recall and accuracy.