基于样例的动画角色多样化研究

发布时间：2018-07-05 02:29

本文选题：动画角色 + 模型融合　；参考：《合肥工业大学》2011年硕士论文

【摘要】：基于样例的快速建模方法一直是计算机3D建模领域的热点研究方向之一,将样例建模引入动画角色三维模型多样化创作及展示中,在动漫影视、3D游戏等领域存在着大量需求。本文在对样例建模研究的广泛调研基础上,针对动画角色模型的特点,分别从模型融合和微变形两个角度探索实现角色模型的多样化生成,研究两种方法各自的应用范围和算法特点,并在此基础上,建立动画角色的模型多样化随机生成系统,最后给出了多样化角色场景的自动生成效果。本文的主要工作包括以下几个方面: (1)针对低精度模型顶点稀疏和广泛使用贴图表现的特点,提出了一种基于简化调和映射的方法提高网格融合的效率,并利用贴图合成技术保留了源模型贴图的内容并生成过渡贴图内容,保证了新生成模型的完备性,在3D模型的数据重用和多样化上具有重要的意义。 (2)根据动画及游戏领域对同类角色多样化的要求,同时满足对时空复杂度的严格限制,提出了基于微变形的多样化模型生成方法。利用参考式变形和路径式变形等微变形方法生成“主体相同,细节各异”的新模型,在保证算法的执行效率和代价基础上,丰富场景中的角色形态。 (3)在前述研究工作的基础上,实现了动画角色的模型多样化随机生成系统。首先,建立参数模式库来存储满足条件的多样化参数模式集合。其次,根据场景需求并辅以随机的方式批量载入参数模式集。最后,根据参数模式中完备的参数信息完成对多样化结果的解析和复现,并自动生成满足要求的多样化动画角色场景。
[Abstract]:The fast modeling method based on sample has been one of the hot research directions in the field of computer 3D modeling. There is a great deal of demand in the field of animation film and television 3D game to introduce sample modeling into the creation and display of 3D model of animation characters. In this paper, based on the extensive research on the modeling of the sample, according to the characteristics of the animation role model, this paper explores the diversification of the role model from the perspective of model fusion and micro-deformation. Based on the study of the application scope and algorithm characteristics of the two methods, a model diversification random generation system for animation roles is established. Finally, the automatic generation effect of the diverse role scenes is given. The main work of this paper includes the following aspects: (1) aiming at the characteristics of low precision model vertex sparse and widely used map representation, a method based on simplified harmonic mapping is proposed to improve the efficiency of mesh fusion. The content of the source model map is preserved and the transitional map content is generated by using the mapping composition technique, which ensures the completeness of the newly generated model. It has important significance in data reuse and diversification of 3D model. (2) according to the requirements of animation and game field for the diversification of similar characters, it meets the strict limitation of space-time complexity. A method for generating diversified models based on micro deformation is proposed. A new model with the same subject and different details is generated by using the methods of reference deformation and path deformation, which can ensure the efficiency and cost of the algorithm. (3) on the basis of the previous research work, the model diversification random generation system of animation characters is realized. First of all, a parameter schema library is established to store a variety of parameter pattern sets that satisfy the conditions. Secondly, the parameter pattern set is loaded in batches according to the requirements of the scene and in a random manner. Finally, according to the complete parameter information in the parameter mode, the analysis and reproduction of the diversified results are completed, and the multi-character scene of the animation is generated automatically.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TP391.41

【参考文献】