Ontology Based Retrieval for Medical Images Using Low Level Feature Extraction

Abstract

This paper presents a system implemented to evaluate the retrieval efficiency of images when they are semantically indexed using a combination of a web ontology language and the low-level features of the image. Images attached to various entities such as online medical report of a patient are abundant on the web but it is difficult to retrieve accurate information from these entities alone, as using entity names in the search engine gives imprecise results especially where images are concerned. To improve the number of results for medical images we introduced Ontology based image retrieval which is a contextual based images search. Along with the ontologies, we have proposed the use of low level feature extraction to identify a particular image and retrieve it as per user’s requirement. The textual and image information is first segregated and relevant data is taken into consideration with the help of semantic filtering based on the contextual exploration (CE), the feature extraction is then applied onto the image and the ontologies are instantiated giving a more relevant result.