Synthesis and Analysis Techniques for the Human Body: R&D Projects - 3D Modeling and Animation: Synthesis and Analysis Techniques for the Human Body - page 343

avoided by representing the posterior distribution non-parametrically. Authors
Bobick & Davis (2001) introduced a real-time human activity recognition
method, which was tested on aerobic exercises. This method is based on a two-
component image representation of motion, the Motion Energy Image, MEI, a
binary image, which displays where motion has occurred during the movement
of the person, and the Motion History Image, MHI, a scalar image, which
indicates the temporal history of motion. MEI and MHI temporal templates are
then matched to store instances of views of known actions.
Recently, in Sminchisescu & Triggs (2001), 3D human motion tracking from
monocular image sequences is achieved by fitting a 3D human body model,
consisting of tampered superellipsoids, on image features (edges and motion
attributes) by means of an iterative cost function optimization scheme. Also,
Plänkers & Fua (2001) present a framework that retains an articulated structure
represented by sticks, but replace the simple geometric primitives by soft objects.
This results in a realistic model where body parts such as the chest, abdomen or
biceps muscles are well modeled.
The main objective of the chapter is to present and analyze the results of
synthesis and analysis techniques for the human body reported by R&D projects
worldwide. Human body synthesis and analysis is a very important research area
with a large number of industrial applications. The examined technological area
has produced impressive research results, which, in many cases, have emerged
as successful consumer applications, especially in the media and film-making
markets. The annual SIGGRAPH Conference is an excellent focal point to
monitor scientific results and their use in several pilot applications. The elimina-
tion of hidden lines in wire-frame renderings, texture mapping, ray-traced
images, animation and expression methodologies are only a few milestones
during recent years in the quest to capture reality. The chapter aims to
demonstrate the incorporation of recent and innovative techniques in human
body modeling, animation and transmission in specific applications developed in
R&D projects worldwide.
The chapter is organized as follows: the next section provides an overview of the
fundamental standards, either established or emerging, which enable the design
and development of interoperable, expandable, reusable and cost-effective
modeling and animation applications. In the section following, a brief presenta-
tion of on-going and state-of-the-art R&D projects in the area of human (or
human parts) analysis and synthesis is presented. This section focuses on
developed or “under development” projects, mainly dealing with transferring
research results to real life applications. The fourth section deals with a detailed
presentation of four recently started European R&D projects that constitute
major applications of the analysis/synthesis technology, developed with the
author's contribution. These sets of applications utilize technologies for 3D