Ontology Based Image Retrieval by Utilizing Model Annotations and Content

Study on image indexing and retrieval concepts have evolved the research further on Content Based Image Retrieval (CBIR). The CBIR system analyses the image content features such as shape, texture, and colour. In recent systems, these low level features are combined with high level semantics in order to achieve high performance. Analyzing the image content will increase the time of feature extraction and complexity of the retrieval process. The proposed system describes the content of an image region using Color and texture as visual features. An ontology model is constructed to analyze the contents. Content Analysis comprises image segmentation and annotation, which scrutinizes images by subdividing them into entities, selecting primary entities, and finally extracting feature signifiers possessed by these entities. To explore the mechanism of content analysis system, a new fused framework using content and model based annotations was proposed. The proposed integrated framework for CBIR system was experimented and the performance was assessed over a large image database with number of features, compared against traditional approaches from the literature. From the experimental results, it is evident that the proposed system performs substantially better and faster than the existing systems.