Content-based object organization for efficient image retrieval in image databases

Much research has focused on content-based image retrieval (CBIR) methods that can be automated in image classification and query processing. In this paper, we propose a blob-centric image retrieval scheme based on the blobworld representation. The blob-centric scheme consists of several newly proposed components, including an image classification method, an image browsing method based on semantic hierarchy of representative blobs, and a blob search method based on multidimensional indexing. We present the database structures and their maintenance algorithms for these components and conduct a performance comparison of three image retrieval methods, the naive method, the representative-blobs method, and the indexed-blobs method. Our quantitative analysis shows significant reduction in query response time by using the representative-blobs method and the indexed-blobs method.