Onto Tagger: Ontology Focused Image Tagging System Incorporating Semantic Deviation Computing and Strategic Set Expansion
Gerard Deepak, Sheeba Priyadarshini · International Journal of Computer Science and business informatics/International journal of computer science and business informatics · 2016
Social Tagging of images uploaded to the Web is highly mandatory as tags serve as the entities for image retrieval. Manual Tagging of images makes the overall process tedious and moreover the tags when manually assigned become noisy. Several automatic tag recommendation systems are available but the background study proves that the tag relevance is not very high. In the era of Semantic Web, there is a need for a semantic driven tagger which would perform efficiently. Also, a system which bridges the gap between manual and automatic tag recommendation is required. An ontology driven semantic tagger for tagging images with social importance which tags the images based on limited reference tags is proposed. The proposed methodology combines ontology crawling using K-Means Clustering and Semantic Deviation Computation using Modified Normalized Google Distance Measured. The tag space is enhanced using Strategic Set Expansion incorporating a dynamic semantic deviation computation. An average precision percentage of 84.4 and an F-Measure percentage of 86.67 are achieved.