Region Based Image Retrieval using k-means and Hierarchical Clustering Algorithms

摘要

Region Based Image Retrieval (RBIR) is an image retrieval approach which focuses on contents from regions of images. This approach applies image segmentation to divide an image into discrete regions, which if the segmentation is ideal, it corresponds to objects. Thus the capture of region is improved so as to enhance the indexing and retrieval performance and also to provide a better similarity distance computation. During image segmentation, a modified k-means algorithm for image retrieval is developed where hierarchical clustering algorithm is used to generate the initial number of clusters and the cluster centres. In addition, during similarity distance computation, object weight based on object’s uniqueness is introduced. Therefore considering images based on regions using RBIR allows the users to pay more attention to regional properties that may better characterize objects which are also made up of local regions. This strategy is able to better reflect the characteristics of the images from the perspective of image regions and objects.