项目名称: 基于动态点云的人脸表情建模和编辑方法研究
项目编号: No.61303088
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 迟静
作者单位: 山东财经大学
项目金额: 26万元
中文摘要: 基于动态点云的表情建模和编辑为三维人脸表情动画的生成提供了新途径,在数字娱乐、虚拟现实、远程网络会议、可视电话等领域有着广泛的应用需求。本项目面向交互式三维人脸表情建模和编辑系统的实际需求,研究基于动态点云的表情建模新方法,重点解决一致点对应关系的建立、表情细节特征的精确匹配、以及建模的自动实现问题,使对动态点云自动建立起高精度且保持点对应一致性的动态模型,动态模型既能精确的再现人脸表情变化,又能为各种后续处理提供强健支持。在此基础上,研究对重建的人脸动态模型进行表情编辑的新方法,重点解决表情合成的真实感问题和用户交互模式的设计问题,为支持高层和低层两种不同的交互编辑方式提供新技术,得到符合用户要求且真实自然的表情。本项目的研究成果将为基于动态点云的真实感三维人脸表情动画制作提供新思路、新理论和新方法,同时也为其它领域的相关问题的解决提供简单实用、稳定有效的新技术。
中文关键词: 动态点云;表情建模;表情编辑;网格对准;表情动画
英文摘要: Using time-varying point clouds for modeling and editing expression is a new way of generating 3D facial expression animation. It is widely applied in many fields such as digital entertainment, virtual reality, teleconference, videophone, and so on. This project is based on the actual demand of interactive 3D facial expression modeling and editing system. We first study expression modeling based on time-varying point clouds and focus on establishing coherent point correspondences, matching expression details, and achieving automated modeling, so as to propose a new method that can automatically build a facial dynamic model with coherent point correspondences from time-varying point clouds. The dynamic model not only accurately reproduces time-varying facial expressions, but also strongly supports subsequent treatments. We also study expression editing based on the reconstructed facial dynamic model and focus on synthesizing high-realistic expressions, and designing friendly user interaction model, so as to provide new techniques for high-level interaction model and low-level interaction model respectively. These techniques can synthesize real and nature expressions satisfying user-defined constraints. The research results of this project will provide new theories and methods for realistic 3D facial expression an
英文关键词: Time-varying point clouds;Expression modeling;Expression Editing;Mesh registration;Facial animation