3D Model-Based Semantic Categorization of Still Image 2D Objects

Automatic classification and interpretation of objects present in 2D images is a key issue for various computer vision applications. In particular, when considering image/video, indexing, and retrieval applications, automatically labeling in a semantically pertinent manner/huge multimedia databases still remains a challenge. This paper examines the issue of still image object categorization. The objective is to associate semantic labels to the 2D objects present in natural images. The principle of the proposed approach consists of exploiting categorized 3D model repositories to identify unknown 2D objects, based on 2D/3D matching techniques. The authors use 2D/3D shape indexing methods, where 3D models are described through a set of 2D views. Experimental results, carried out on both MPEG-7 and Princeton 3D models databases, show recognition rates of up to 89.2%.

3D Model-Based Semantic Categorization of Still Image 2D Objects

Automatic classification and interpretation of objects present in 2D images is a key issue for various computer vision applications. In particular, when considering image/video, indexing, and retrieval applications, automatically labeling in a semantically pertinent manner/huge multimedia databases still remains a challenge. This paper examines the issue of still image object categorization. The objective is to associate semantic labels to the 2D objects present in natural images. The principle of the proposed approach consists of exploiting categorized 3D model repositories to identify unknown 2D objects, based on 2D/3D matching techniques. The authors use 2D/3D shape indexing methods, where 3D models are described through a set of 2D views. Experimental results, carried out on both MPEG-7 and Princeton 3D models databases, show recognition rates of up to 89.2%.

Mining 3D Shape Data for Morphometric Pattern Discovery

Recent technological advances in 3D digitizing, noninvasive scanning, and interactive authoring have resulted in an explosive growth of 3D models in the digital world. There is a critical need to develop new 3D data mining techniques for facilitating the indexing, retrieval, clustering, comparison, and analysis of large collections of 3D models. These approaches will have important impacts in numerous applications including multimedia databases and mining, industrial design, biomedical imaging, bioinformatics, computer vision, and graphics. For example, in similarity search, new shape indexing schemes (e.g. (Funkhouser et al., 2003)) are studied for retrieving similar objects from databases of 3D models. These shape indices are designed to be quick to compute, concise to store, and easy to index, and so they are often relatively compact. In computer vision and medical imaging, more powerful shape descriptors are developed for morphometric pattern discovery (e.g., (Bookstein, 1997; Cootes, Taylor, Cooper, & Graham, 1995; Gerig, Styner, Jones, Weinberger, & Lieberman, 2001; Styner, Gerig, Lieberman, Jones, & Weinberger, 2003)) that aims to detect or localize shape changes between groups of 3D objects. This chapter describes a general shape-based 3D data mining framework for morphometric pattern discovery.