CASIA OpenIR  > 毕业生  > 博士学位论文
基于物理及数据驱动的特效动画模拟
其他题名Physically Based and Data Driven Effects Animation Simulation
徐士彪
2014-05-30
学位类型工学博士
中文摘要随着计算机动画技术的飞速发展,真实感特效动画的模拟和互动呈现逐渐成为动画创作中的重要而 关键的精髓内容,其中,最具有挑战性的流体及表情动画作为典型的特效动画备受关注。对于复杂的流体效果,只有借助较为精确的物理方程才能再现其自身的运动 规律及外在的视觉特效,现阶段所能够模拟的流体现象无论是从真实感、实时性,还是从复杂程度上都有了显著的提高。但是对于细节丰富而又广泛存在于现实世界 中的流体现象,现有流体模拟技术在数值实时精确计算和效果真实感渲染等方面都还不尽人意。对于表情动画的模拟,要求相比于其他类型特效动画更高,现阶段的 表情动画模拟技术虽然能够快速的完成计算,灵活的生成真实的表情动画,但仍存在诸多问题,同时表情动画的真实感渲染,相关的研究工作较少,真实感效果较差。 在应用方面,无论是影视特效、视频游戏,还是互动应用,人们对于高质量、高细节流体特效的实时模拟及真实感表情动画的模拟与渲染需求也十分强烈。 因此,本文的工作基于上述研究背景展开,致力于研究更好的特效模拟方法以及更佳的特效呈现效果。本文的研究将对提高特效动画制作的质量与效率、增加实时体 感互动体验的真实性与沉浸性有着重要的理论和现实意义。 本文基于物理及数据驱动的方式,对真实感特效动画的模拟进行了有针对性的研究,论文的贡献主要体现在以下几个方面: 第一,提出真实感水墨动态扩散效果的实时模拟方法,针对水墨扩散现象,基于水墨流体的组成成分提出一种真实、高效的物理模型:水墨的网格粒子混合 (Grid-Particle representation)形态,并基于图形硬件进行解算,克服传统方法引入过多的模型与假设、计算复杂等特点; 第二,提出一种真实感水滴自由散落效果的实时模拟方法,对通用Smoothed Particle Hydrodynamics (SPH)方法进行改进,提出一种精确SPH模型来模拟小比例的水滴现象,能够精确的维持流体线动量及角动量的守恒,同时基于新一代着色器模型,提出一种 适用于水滴渲染的光线追踪方法,克服了传统流体渲染独立于摄像机、计算量大等特点; 第三,提出了一种实时表情动画模拟与真实感渲染方法,表情动画的模拟主要包括表情运动单元的简化、基于统计训 练的表情图像半自动标定、旋转不变的人脸检测方法及基本表情权重值的自动计算与映射,能够显著提高表情动画模拟的精度与效率;表情动画的真实感渲染通过将 基于物理的光线漫反射分布及镜面反射分布相结合,并采用预计算方法进行渲染,显著提高渲染的真实感与效率。
英文摘要With the rapid development of computer animation technology, realistic effects animation simulation and interactive rendering become the important and key contents in animation creation gradually, in which, the most challenging fluid and facial animation as the typical effects animation have attracted more attention. For complex fluid effect, only the accurate physical equation can reproduce the motion law and the external visual effects, at present, fluid simulation techniques have been improved significantly in terms of realistic and efficient. However, for the rich details in the fluid phenomena in the real world, the existing numerical fluid simulation technologies are unsatisfactory in aspects of precise calculation and realistic rendering. For facial animation simulation, the requirement is higher than other types of special effects animation, at present, facial animation simulation technology is able to complete the calculation quickly and the production of flexible real expression animation, however, there are still many problems. At the same time, the relevant research work of realistic rendering of expression animation is less, and the realistic effect is poor. In the aspect of application, the film and television special effects, video games, and interactive applications require high quality fluid effects in real-time and realistic facial animation. Therefore, based on the above research background, this paper dedicates to study better simulation method and the special effects rendering effect. In this paper, the purpose is to improve the quality and efficiency of the special effects animation, increase realistic feeling in interactive experience, which is important in theoretical and practical creation. Based on the physical and data driven method, this paper focuses on the simulation of realistic effects animation. The main contribution of the dissertation is listed as follows: First, we present an effective method to simulate the ink diffusion process in real time that yields realistic visual effects. Our algorithm updates the dynamic ink volume using a hybrid grid-particle method. Compared to traditional fluid simulation methods that treat water and ink as two mixable fluids, our method is simple but effective. The simulation and the rendering processes are efficiently implemented on graphics hardware at interactive frame rates. Second, we present an interactive simulation method for water drop effects. Our algorithm takes an accurate SP...
关键词特效动画 水墨扩散 水滴散落 表情动画 Effects Animation Ink Diffusion Water Droplets Facial Animation
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/6649
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
徐士彪. 基于物理及数据驱动的特效动画模拟[D]. 中国科学院自动化研究所. 中国科学院大学,2014.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
CASIA_20111801462909(14234KB) 暂不开放CC BY-NC-SA
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[徐士彪]的文章
百度学术
百度学术中相似的文章
[徐士彪]的文章
必应学术
必应学术中相似的文章
[徐士彪]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。