3D Object Retrieval

Abstract

In the last few years, the various technologies related to acquiring 3D models of everyday objects has seen a steady increase. These models are then being used in many areas such as, engineering, museums, cinematics and healthcare. Hench, there has been an increasing need for 3D content-based object retrieval and it has become the focus of many research activities. One of the important research area is to find sufficient number of descriptors in order to capture local and global characteristics of an object. In this paper, we propose joint learning of the view-model relevance for retrieval of 3D objects by formulating the objects using different structures of graph such as hyper graphs. Here the model-based and view-based relevance are jointly explored in a graph-based framework.