Abstract
Conference Title: 2019 International Conference on Computer and Information Sciences (ICCIS) Conference Start Date: 2019, April 3 Conference End Date: 2019, April 4 Conference Location: Sakaka, Saudi Arabia Due to the subjectivity nature of social tagging, noisy and incompletion of human generated tags, an automatic image annotation system is required to filter and expand the associated tags for efficient image retrieval in social photo web sites. This paper presents a Hybrid based semantic and context image annotation system for noise tag filtration and tag expansion. Experimental results shows that, the proposed system effectively identify and expand tags and enhance performance of image retrieval in social photo web sites.