An Efficient Method of Improving Image Retrieval Using Combined Global and Local Features
Abderrahim Khatabi, Amal Tmiri, Ahmed Serhir · Lecture notes in electrical engineering · 2016
Nowadays, with the increased use of digital images it has become essential to find an efficient system for searching and indexing of images from large image collections. CBIR systems can be used for searching and retrieving different kinds of images from large databases on the bases of the visual content of the images. Currently, CBIR techniques work on combination of low level features i.e. color, shape and texture. In this paper we have designed a content based image retrieval system based on the combination of local and global features. The local features are obtained through local binary pattern (LBP) technique which is used to extract texture-based features from an image, while the global features are extracted using Angular Radial Transform (ART). To demonstrate the efficacy of this combination, experiments are conducted on Columbia Object Image Li-brary (COIL-100) and MPEG-7 shape-1 part B database. The result showed significant improvement in the retrieval accuracy when compared to the existing system.