SemFace: A Human Face Search Engine Based on the Semantic Web

SemFace is Semantic Web based search engine to retrieve human face images through face facial features. Each human face image has two ontologies (internal and external) that are associated with it. The classification of those ontologies is based on the data which are concerned for each image. Internal ontology is concerned with absolute (deterministic) data, whereas external ontology is concerned with relative (nondeterministic) data. This search engine consists of two steps (or more) to retrieve a human face images. The first step uses a user query that is the description of some face in order to. retrieve the image most similar to that face. The second step is based on the semantic similarity between the selected images from the first step with other images. To mitigate the differences of interpretations between annotators and retrievers, we use coarse values for human facial attributes. One of the most promising issue that this work proposes is that the SemFace may be used in multimedia applications as well as increasing the efficiency of low image retrieval algorithms.