Iterative Region Merging and Object Retrieval Method Using Mean Shift Segmentation and Flood Fill Algorithm

Abstract

Nowadays image retrieval plays an important role in an extraordinary number of multimedia applications which serve human society. Since a generic graph based retrieval scheme working in every situation can be quite difficult to implement due to the presence of complex scene. In this paper we investigate to deal with such problems and focuses on how to extract the object from an image using topological models. It begins with the innovative concept of initial low level segmentation of input image, which are used to construct topological model on the basis of connected regions. Using this topological model we have design a new prototype known as Iteratively Region Merging and Object Retrieval (IRMOR). Using this prototype IRMOR we can extracted contour based descriptors on an object from an image which constituent the high level features. In addition we focus on how to apply a region labeling and flood fill method to extract object after formation of object contour, IRMOR have special application such an image matching, object recognition, content based image retrieval, object tracking etc.