真实感三维头像的设计

发布时间：2018-01-28 04:10

本文关键词： 图像识别建模技术细分技术贴图技术　出处：《河海大学》2007年硕士论文　论文类型：学位论文

【摘要】： 近年来，随着大众文化娱乐的日趋多元化以及数码特效技术的不断创新，动漫文化得以开始繁荣，出现了三维动画、FLASH动画、全息动画等崭新的动漫形式。三维动画作为一项新型的产业已经具备相当庞大的规模。但三维动画制作在国内尚处于起步阶段，与国外一些三维动画制作强国相比还有较大差距。在三维动画制作过程中，存在很多问题。首先在硬件方面，要做电脑动画离不开高性能的电脑。虽然现在PC在速度及存储量上都有显著的提高，但是图形工作站在速度、存储容量等方面还是比一般的PC机优越。其次在软件方面，现在许多角色设计都集中在两足动物的步伐设计上。现有的头部角色设计一般是使用几个简单的角色库，用户只能通过调用库中已有的头像动作，来完成自己的设计，很难自行设计头像的动作。这样，人物头部的设计就过于简单。因此，三维头像的设计就显得十分重要。本文采用基于扫描线的人脸图像识别技术对人脸图像进行识别，通过建立活性边表存储识别的数据，提高识别效率，以提取的人脸图像的轮廓图作为头像建模的基础；采用三维头像建模技术对提取出的轮廓图进行三维建模，提出了从正、侧面照片中提取面部特征的局部曲率最大跟踪算法定义脸部的特征点，利用特征点的偏移量计算非特征点的偏移量来合成脸部模型点的三维坐标；采用基于Hemite样条插补的三维网格曲面的细分方法对生成的简化三维头像进行细分，生成真实感强的三维头像；采用纹理映射技术，对生成的三维头像进行贴图，，使三维头像更加逼真。应用该方法进行三维头像的设计，既可以大大减少设计制作的工作量，又可以方便的得到所需的真实感三维头像。
[Abstract]:In recent years, with the increasing diversity of popular cultural entertainment and the constant innovation of digital special effects technology, animation culture has begun to prosper, and 3D animation flash animation has emerged. Holographic animation and other new forms of animation. 3D animation as a new industry has a very large scale, but the production of 3D animation in China is still in its infancy. Compared with some foreign countries, there is still a big gap. In the process of 3D animation production, there are many problems. First of all, in the hardware, computer animation can not do without high-performance computer, although the PC has significantly improved in speed and memory. But the graphics workstation is superior to the normal PC in speed, storage capacity and so on, followed by the software. Nowadays, many character designs are focused on the design of bipods. The existing head character design generally uses several simple role libraries, and users can only call the avatar action in the library. To complete their own design, it is difficult to design their own movements of the avatar. In this way, the design of the head of the character is too simple. Therefore, the design of the three-dimensional image is very important. In this paper, the face image recognition technology based on scan line is used to recognize the face image, and the recognition efficiency is improved by establishing the active edge table to store the recognition data. The contour image of the extracted face image is used as the basis of the head image modeling. Using 3D head image modeling technology to model the extracted contour image, a local curvature maximum tracking algorithm for extracting facial features from positive and lateral photographs is proposed to define the feature points of the face. The 3D coordinates of facial model points are synthesized by calculating the offsets of non-feature points by using the offsets of feature points. The subdivision method of 3D mesh surface based on Hemite spline interpolation is used to subdivide the generated simplified 3D head image to generate a realistic 3D head image. Texture mapping technology is used to map the generated 3D head image to make the 3D head image more realistic. Using this method to design 3D head image can greatly reduce the workload of design and production. It is convenient to get the realistic 3D image.
【学位授予单位】：河海大学
【学位级别】：硕士
【学位授予年份】：2007
【分类号】：TP391.41

【引证文献】