High-level Shape Representations for Content Creation

我们建议研究高级表示以促进内容创建。几何内容的创建在计算机图形和动画、视觉效果、游戏开发和面向制造的设计等领域是一项重要任务。创建3D内容最常见的方法是直接对3D形状的几何形状进行建模。然而，建模是一项艰巨的任务，艺术家需要熟练使用传统的软件工具，这通常涉及在低级形状表示，例如，三角形网格工作。因此，在过去的十年中，计算机图形学的努力集中在促进3D模型创建的任务上。这个方向最近的一个趋势是将形状表示为在更语义层次上定义的原语集合，例如形状部分（如椅子的腿、座位和靠背）。这些表示允许在更抽象的级别上操作形状，独立于底层的低级表示。高级表示还可以用于合成新模型，或通过捕获形状结构和外观的其他方面来指导形状建模或探索形状空间。在此背景下，提出的研究计划的目标是为形状操作和合成开发高级表示，以促进几何内容的创建，特别是3D形状。这种表示的成功使用需要三个组成部分：1)适合目标应用的形状表示；2)自动或半自动建立已有几何数据表示的算法；3)引导机制，允许用户使用表示进行内容创建。因此，这三个组成部分组成了该计划的研究流。在第一个流中，我们将研究更适合操作和合成的形状表示的设计。在第二个流中，我们的目标是研究大数据分析如何在形状分析方面取得重大进展，并促进高级表示的自动构建。在第三流，我们将调查指导机制的发展，可以支持创造性的建模。我们希望，从长远来看，这个项目可以促进内容的创作，从而引起设计和娱乐行业的兴趣。

We propose to investigate high-level representations to facilitate content creation. The creation of geometric content is an important task in areas such as computer graphics and animation, visual effects, game development, and fabrication-oriented design. The most common approach for creating 3D content is by directly modeling the geometry of 3D shapes. However, modeling is a laborious task where artists need to be skilled to use the conventional software tools, which typically involve working on low-level shape representations, e.g., triangle meshes. ***Thus, during the last decade, efforts in computer graphics have focused on the task of facilitating the creation of 3D models. A recent trend towards this direction is to represent shapes as a collection of primitives defined at a more semantic level, e.g., shape parts (such as the legs, seat and back of a chair). These representations allow manipulating shapes at a more abstract level, independently of the underlying low-level representation. High-level representations can also be used to synthesize new models or to guide the modeling of shapes or exploration of a shape space by capturing additional aspects of the structure and appearance of shapes.***In this context, the goal of the proposed research program is to develop high-level representations for shape manipulation and synthesis to facilitate the creation of geometric content, especially 3D shapes. The successful use of such representations requires three components: 1) A shape representation that is suitable for the target applications; 2) Algorithms that automatically or semi-automatically build the representations for existing geometric data; and 3) Guiding mechanisms that allow a user to employ the representations for content creation. Thus, these three components comprise the research streams of the program. ***In the first stream, we will investigate the design of shape representations that are more suitable for manipulation and synthesis. In the second stream, our goal is to study how big data analytics can lead to significant advancements in shape analysis, and facilitate the automatic construction of the high-level representations. In the third stream, we will investigate the development of guiding mechanisms that can support creative modeling. We expect that, in the long term, this program can contribute to facilitating the creation of content and thus be of interest to the design and entertainment industries.