基于手绘输入理解的智能3D几何建模系统

近年来,图形学领域已经开发了大量的3D建模和可视化工具,但是这些工具的界面往往过于复杂,制约了设计人员的创造性工作,也进一步限制了这些工具为普遍用户所使用。事实上,理想的3D建模工具应该提供给用户以尽可能自然、简单的操作,例如基于电子笔的草图输入,而且随着移动设备的发展,也使之成为图形学中更有前景的研究方向。文章论述了当前支持手绘的建模技术的研究现状,并提出了一个更加智能化的多模块建模体系,突出了交互技术、智能理解和高质量的可视化输出技术的结合;而且针对手绘输入的不明确性,着重分析了图形数据的基于特性的语义描述,提出了智能的逻辑推理模型,并进一步描述了下一代融合自然输入、逻辑推理与虚拟显示于一体的智能图形建模系统的实现。预测了人工智能技术与CA D技术的结合将是未来图形学领域发展的主要方向之一。

Sketch-based Intelligent 3D Geometric Modeling System

The past research in the field of 3D computer graphics brought to life a large amount of 3D modeling and rendering .tools.The user interface of these tools is however very complicated to learn and use,which leaves 3D graphics out of reach of majority of users.Sophisticated user interfaces that aspire to hide the complexity of underlying software, such as sketch-based interfaces,are starting to gain popularity.And the spreading of mobile computers for which sketching serves as a main means of input makes tbe new interfaces even more interesting.In this paper we devote to research the modeling technology based on sketching.We will present a muhimodal modeling system based on sketching input,and state a series of specific implementation.Furthermore,aiming at the ambiguity of free form drawing,we focus on the semantic representation of shape and the intelligent reasoning technique,and describe the integrated process from nature input,logic reasoning to visual rendering;in the end we foresee the combination of AI and CAD is one promising topic of future computer graphics researches.