Ontology-Based Image Retrieval with SIFT Features

A new ontology-based image retrieval framework which brings in SIFT features is presented in this paper. Firstly, it brings SIFT features into the image ontology to build an ontology library which describes the image element together with shape, color, texture and other low-level features. And then, calculate the similarity of SIFT features, low-level features, concept of ontology and semantic features between the sample images and the image elements in the ontology library to retrieve images. The image semantic is taken into account in the framework which also using the SIFT features to maintain rotation invariance and scale invariance. Therefore, it can improve the accuracy of image retrieval.