欢迎访问《图学学报》 分享到:

图学学报

• 交互设计与虚拟现实 • 上一篇    下一篇

人群仿真中角色纹理多样化方法

摘 要:人群仿真中角色模型的多样化能营造出符合实际的效果。通过模型纹理处理技术来加强人群仿真中的人物外观多样化,是解决仿真场景中三维模型外观差异性的有效手段。提出一种基于模板的纹理多样化生成方法。首先给不带纹理坐标的网格模型展UV后作为纹理模板;然后利用基于图像分割方法处理UV展开图得到纹理的衣饰语义信息;在此基础上实现纹理图像的纯色着色和图案填充,从而生成各种不同的纹理,实现人群仿真的角色多样化。所提方法实用可行,较好地解决了虚拟角色制作效率问题,快速生成视觉上合理的大规模人群。#br# 关 键 词:纹理;UV;多样化;人群仿真   

  • 出版日期:2014-02-28 发布日期:2015-03-26

Texture Diversity Generating Method for Characters in Crowd Simulation

Abstract: Diversity appearance of character can create a realistic effect. It is an effective way to resolve the difference of the character appearance in crowd simulation scenarios by strengthening the diversity appearance of characters with texture. This paper puts forward a method to achieve diversification of the appearance by texture templates. First, UV is unfolded for the mesh model without texture coordinates as texture templates. Second, the texture templates are processed to get the semantic information of clothing based on image segmentation. Finally, the texture templates are filled with pure color or image tiles, and various textures can be gotten to achieve the goal of diversity appearances in crowd simulation. The proposed method is practical and feasible to solve creation efficiency issues of virtual character, and can quickly generate visually-convincing large-crowd simulation results.#br# Key words: texture; UV; diversification; crowd simulation   

  • Online:2014-02-28 Published:2015-03-26

摘要: 人群仿真中角色模型的多样化能营造出符合实际的效果。通过模型纹理处理技术来加强人群仿真中的人物外观多样化,是解决仿真场景中三维模型外观差异性的有效手段。提出一种基于模板的纹理多样化生成方法。首先给不带纹理坐标的网格模型展UV后作为纹理模板;然后利用基于图像分割方法处理UV展开图得到纹理的衣饰语义信息;在此基础上实现纹理图像的纯色着色和图案填充,从而生成各种不同的纹理,实现人群仿真的角色多样化。所提方法实用可行,较好地解决了虚拟角色制作效率问题,快速生成视觉上合理的大规模人群。

关键词: 纹理, UV, 多样化, 人群仿真

Abstract: Diversity appearance of character can create a realistic effect. It is an effective way to resolve the difference of the character appearance in crowd simulation scenarios by strengthening the diversity appearance of characters with texture. This paper puts forward a method to achieve diversification of the appearance by texture templates. First, UV is unfolded for the mesh model without texture coordinates as texture templates. Second, the texture templates are processed to get the semantic information of clothing based on image segmentation. Finally, the texture templates are filled with pure color or image tiles, and various textures can be gotten to achieve the goal of diversity appearances in crowd simulation. The proposed method is practical and feasible to solve creation efficiency issues of virtual character, and can quickly generate visually-convincing large-crowd simulation results.

Key words: texture, UV, diversification, crowd simulation