CAREER: Motion Transformations for Animation

点击翻译按钮获取中文摘要

Computer Animation is a powerful medium for communication, with potential to explain, educate, orentertain. However, the potential of animation beyond entertainment is largely untapped because of thedifficulty and cost in producing it. While the state of the art provides tools for expert animators to craftworks of art, it offers little help for those without considerable time and talent.Our vision is that computer animation will be a widely used medium for communication, that a wide rangeof users will choose to share their messages in this medium, and that the availability of this medium willimpact how we educate students. The widespread availability of internet bandwidth offers the mechanismfor creators to share animation: what is missing are the tools to create it.Our premise is that:1. The power of animation can be more broadly used, with particular impact in education, if we can lowerthe barriers to its production with tools that democratize the medium.2. Techniques for transforming motions can enable such tools by allowing users to leverage the work ofothers if we can devise technologies that retain the desirable properties of the original duringtransformation.3. The spacetime constraints approach to motion transformation can provide this technology if it is morefully developed.This proposal presents a faculty career development plan committed to the vision of ubiquitous animation.The research component will develop technologies that enable a wider range of uses of animation, and theeducation component will use animation for education and better prepare computer scientists to work withvisual media.The overall goal of the research effort is to create enabling technologies for new animation tools that aremore accessible to those who are not skilled animators. We propose techniques for transforming motions tonew uses, for example applying an action to a new character or new environment. The challenge is to createtransformations that both meet the new needs, yet preserve the desirable properties of the original. Our prior work introduced the spacetime constraints approach to motion transformations. Mathematically, theapproach poses motion transformation as a constrained optimization problem: what motion meets thespecified needs yet best preserves the original motion? Unlike most animation techniques, spacetimeconstraints consider entire motions simultaneously, rather than computing instants in time individually. This requires solving a single, large numerical problem for each motion computation. Our work to date shows the spacetime constraints approach is a viable approach for tackling important problems in computer animation. The research component of the development plan is to leverage our initial results with the spacetime constraints approach towards the broader goal of democratizing animation. Specifically, work is proposed to improve the mathematical foundations of the approach to enhance its reliability and applicability, to develop new types of motion transformations, to implement systems that provide motion transformations to our target users, and to demonstrate how these new types of tools can be used for the production of effectivevisual messages.The education component of the development plan consists of three parts. First, the plan addresses the needto train students at the University of Wisconsin in graphics and animation by developing new courses andproviding research and project opportunities for students at all levels. Second, the plan proposes to useanimation to improve pedagogy: by creating materials for classes as well as for outreach. Third, the planproposes to help foster the development of animation technology through outreach to the technicalcommunity.