Study On The Elastic Surface Layered Model For 3D Human Body

With the advent of 3D epoch in computer science, 3D human body modeling has been receiving increasing attention from researchers in the fields of image processing and computer vision in the recent years. It has a lot of applications in movie animation, virtue reality and industry design, etc. The human body model evolution from wireframe hierarchies to layered structures is the most important achievement. The modeling methods have covered the knowledge of 3D formation, motion detective, mathematic analysis, biomechanics and anatomy, etc. They are becoming the hotspot of the studies of the experts in computer graphics. There are two characteristics in the relevant studies: the first is that emphasis is on the construction of mathematic model lack of practical body data so that only local parts can be simulated; the other is that there are huge drain on resources to produce the perfect virtual human. There exist only a few investigations on the body models used in 3D movie.The 3D Laser Color Scanning System (3DLCS) developed by domestic researchers is applicable and much body information has been acquired by the system. On the basis of the early work, elastic surface layered model of human body is lucubrated in this dissertation. From outer to the inner side, the layered model can be divided into several layers: skin, muscle, tissue and skeleton. Each one has its characteristic physical attributions so that appropriate methods are required for them. At last the layers need to be integrated to create the 3D human body that has realistic appearance and natural motion.The advantage of layered model lies in the separation of the two problems: skeleton animation and skin deformation. Then the layered character animation pipeline is constructed. The different characteristics of the different anatomical layers suggest a hybrid model in which different modeling techniques are used at each stage of the pipeline to form the final layered model. At the same time, the environmental force and the restriction are studied in details. The canonical equations of motion can be solved by discrete methods.In the process, each layer is modeled by the appropriate approach. The skin layer is expressed as the generalized elastic surface model. The visco-elastic property of skin is discussed to simulate the permanent (aged) and temporal (expressional) wrinkles. On the basis of simple face model, the affection of the muscle mask and tissue on the facial expression is analyzed. The weighting factors for elasticity, viscosity and plasticity are introduced so that they can be varied according to the experimental data to produce facial animation.The skeleton and tissue layers are highly correlated to the motion of muscle. The skeleton layer can be regarded as linked solid architecture. The links are used to represent joints and the skeletal muscles draw them. The tissue layer is treated as the media that transfer the muscle force and can be simulated by mass-spring system. In the system, muscle patches and tissue are simulated by spring grid and angel spring, respectively. Furthermore, Free Form Deformation (FFD) is applied to simplify the deformation of tissue and show the constraints on muscle.Muscle layer is the most complicated one and the emphasis in the dissertation. Firstly, the preprocessed approach is similar to the anatomical one. The relation between surface of muscle and underlying structure is studied in order to comprehend the contraction of muscle. Then the relation between joints and skeletal muscles, the connection between deformation of belly and muscle force are analyzed to help the modeling of skeletal muscles. And the procedural modeling and animation language is used to represent those.In order to satisfy the request of real-time deformation, the modeling of muscle is based on the simple geometrical theory at first. Then the anatomy-based deformation is calculated. There are two steps: the definition of control axis, and the designation of the cross-sections. The binding of axial curve with cross-section an...