A long-standing goal in computer graphics is to create and control realistic motion for virtual human characters. Despite the progress made over the last decade, it remains challenging to design a system that allows a random user to intuitively create and control life-like human motions. This dissertation focuses on exploring theory, algorithms and applications that enable novice users to quickly and easily create and control natural-looking motions, including both full-body movement and hand articulations, for human characters. More specifically, the goals of this research are: (1) to investigate generative statistical models and physics-based dynamic models to precisely predict how humans move and (2) to demonstrate the utility of our motion models in a wide range of applications including motion analysis, synthesis, editing and acquisition. We have developed two novel generative statistical models from prerecorded motion data and show their promising applications in real time motion editing, online motion control, offline animation design, and motion data processing. In addition, we have explored how to model subtle contact phenomena for dexterous hand grasping and manipulation using physics-based dynamic models. We show for the first time how to capture physically realistic hand manipulation data from ambiguous image data obtained by video cameras. The electronic version of this dissertation is accessible from http://hdl.handle.net/1969.1/149245

This book constitutes the thoroughly refereed post-proceedings of the 2004 International Workshop on Intuitive Human Interfaces for Organizing and Accessing Intellectual Assets, held in Dagstuhl Castle, Germany in March 2004. The 17 revised full papers presented together with an introductory overview have gone through two rounds of reviewing and revision. The papers are organized in topical sections on man-machine interface for intuitive knowledge access, intelligent pad and meme media, visualization and design of information access spaces, and semantics and narrative organization and access of knowledge.

This book focuses on two issues related to human figures: realtime dynamics computation and interactive motion generation. In spite of the growing interest in human figures as both physical robots and virtual characters, standard algorithms and tools for their kinematics and dynamics computation have not been investigated very much. "Simulating and Generating Motions of Human Figures" presents original algorithms to simulate, analyze, generate and control motions of human figures, all focusing on realtime and interactive computation. The book provides both practical methods for contact/collision simulation essential for the simulation of humanoid robots and virtual characters and a general framework for online, interactive motion generation of human figures based on the dynamics simulation algorithms.

Abstract: "Automatic generation of realistic human motion has been a long-standing problem in computer graphics. This thesis introduces a novel algorithm for transforming character animation sequences that preserves essential physical properties of the motion. The algorithm maintains realism of the original motion sequence without sacrificing the full user control of the editing process. We use the spacetime constraints dynamics formulation to manipulate the motion sequence. In contrast to most physically based animation techniques that synthesize motion from scratch, we take the approach of motion transformation as the underlying paradigm for generating computer animations. In doing so, we combine the expressive richness of the input animation sequence with the controllability of spacetime optimization to create a wide range of realistic character animations. The spacetime dynamics formulation also allows editing of intuitive high-level motion concepts such as the time and placement of footprints, length and mass of various extremities, joint arrangement and gravity. Our algorithm permits the reuse of highly-detailed captured motion animations. In addition, we describe a new methodology for mapping a motion to/from characters with drastically different number of degrees of freedom. We use this method to reduce the complexity of the spacetime optimization problems. Furthermore, our approach provides a paradigm for controlling complex dynamic and kinematic systems with simpler ones."

This book discusses the concepts, theory, and core technologies of intelligent theory and human animation, including video based human animation and intelligent technology of motion data management and reusing. It introduces systems developed to demonstrate the technologies of video based animation. Lively pictures and demos throughout the text help make the theory and technologies more accessible to readers.

The recent development of motion capture technology and its popularity makes it possible to produce realistic animation from lively captured motions of an actor with much less effort. With these changes in expressive character animation, an effective motion editing approach is required for modifying the prerecorded motion clips since such data are typically specialized to a particular character and conditions at the time of capture and difficult to edit into a desired output. Thus, recent research in the character animation field has focused on techniques or tools for reusing a set of motion capture data as example motions and producing stylistic variations from the set.For effective editing, each motion property can be classified into two categories, qualitative and quantitative aspects of motions. For example, qualitative properties represent the physical values that can be computationally estimated from the motion such as speed, orientation, and forces while quantitative properties are more associated with the abstract description of motions including physical and inner states of a character. Over years, many researches have focused on each aspect for synthesizing stylistic variations from input motion. Only a few works consider both aspects, and then in a limited way such that they either lack flexibility in controls for stylizing the input motion or the high-quality seen in motion capture data. Motivated by this, our goal is to construct a comprehensive motion framework, especially for human locomotion, that adopts motion capture data as our main parameter source and allows an animator to apply a wide variety of stylistic changes via graphical user interfaces (GUIs). We expect that our framework overcomes many limitations in previous systems by editing the quantitative and qualitative aspects of motions with multiple animation systems, where each of them focuses on editing different motion properties during the synthesis process: interactivity, composition, and timing.In interactive editing of locomotion style, our system is particularly designed for making stylistic changes via the extracted correlations between the end effectors, that we name drives, and the body movement. When an animator interactively controls the positional data for the wrists, ankles, and center of mass, the system automatically updates the current pose at each frame based on the driver positions and driven orientations using the inverse kinematic (IK) and balance maintaining routines. The overall editing process is controlled by a set of simple and intuitive linear operations on the motion drives or extracted correlations. Thus, an animator can quickly transform the input locomotion into a desired style at interactive speed.For expressive locomotion generation on an arbitrary path, we provide a system that adopts multiple example clips on a motion path specified by an animator. Significantly, the system only requires a single example of straight-path locomotion for each style modeled and can produce output locomotion for an arbitrary path with arbitrary transitions. Several techniques are applied to automate the overall synthesis: detection of multiple foot-plants from unlabeled examples, estimation of an adaptive blending length for a natural style change, and a post-processing step for enhancing the physical realism of the output animation. Compared with previous approaches, our system requires significantly less data and manual labor, while supporting a large range of styles.When generating locomotion, it is particularly challenging to adjust the motion's style in a qualitative way. The component-based system is designed for human locomotion composition that drives off a set of example locomotion clips. The distinctive style of each example is analyzed in the form of sub-motion components decomposed from separate body parts via independent component analysis (ICA). During the synthesis process, we use these components as combinatorial ingredients to generate new locomotion sequences that are stylistically different from the example set. Our system is designed for any animator who may not have much knowledge of important locomotion properties, such as the correlations throughout the body. Thus, the proposed system analyzes the examples in a unsupervised manner and synthesizes an output locomotion from a small number of control parameters. Our experimental results show that the system can generate physically plausible locomotion in a desired style at interactive speed.Timing plays an important role in specifying how a character moves from one pose to another. To effectively capture the timing variations in the example set and to utilize them for style transfer, we propose an editing system that provides separate controls over temporal properties of an input motion via the global and upper-body timing transfers. The global timing transfer focuses on matching the input motion to the body speed of the selected example motion and will contain the overall sense of emotional or physical state observed in the example. On the other hand, the timing transfer in the upper body propagates the sense of movement flow through the torso and arms, which is often referred to as succession. We try to transfer this succession by capturing the relative changes of angle rotation in the upper body joints from the example motion and then apply them to the input motion with a scaled amount. Overall, this system provides an animator temporal edits on locomotion style without destroying spatial details and constraints preserved in the original motion.

This book constitutes the refereed proceedings of the 9th International Conference on Articulated Motion and Deformable Objects, AMDO 2016, held in Palma de Mallorca, Spain, in July 2016.The 20 papers presented were carefully reviewed and selected from 34 submissions. The conference dealt with the following topics: advanced computer graphics and immersive videogames; human modeling and animation; human motion analysis and tracking; 3D human reconstruction and recognition; multimodal user interaction and applications; ubiquitous and social computing; design tools; input technology; programming user interfaces; 3D medical deformable models and visualization; deep learning methods for computer vision and graphics; multibiometric.