3D Shape Analysis

9Examples of 3D Semantic Applications

9.1 Introduction

The goal of this chapter is to illustrate the range of applications involving 3D data that have been annotated with some sort of meaning (i.e. semantics or label). There are many different semantics associated with 3D data, and some applications consider more than one category of labels, However, many of the current applications lie in the four following categories:

• Shape or status: What is the shape of the 3D object, and whether this is a normal shape or not (see Section 9.2).
• Class or identity: What sort of a 3D object is being observed, or even which specific 3D object (e.g. person) (See Section 9.3).
• Behavior: What are the 3D objects doing (see Section 9.4).
• Position: Where are the 3D objects (see Section 9.5).

This chapter briefly introduces examples of all of these, and the following chapters go more deeply into four of the applications: 3D face recognition (Chapter 10), object recognition in 3D scenes (Chapter 11), 3D shape retrieval (Chapter 12), and cross domain image retrieval (Chapter 13).

9.2 Semantics: Shape or Status

In this section, we consider applications where the 3D data are analyzed to extract usable shapes or properties from the 3D data. A key advantage of having 3D data is the ability to measure real 3D scene properties, such as the sizes of objects or the distance between locations. 2D images allow measurement of image properties (such as the distance between two pixels), which may allow estimation ...