真实感三维人脸建模及应用研究

发布时间：2018-06-21 20:26

本文选题：人脸建模 + 形变模型　；参考：《北京工业大学》2004年博士论文

【摘要】：真实感三维人脸建模和人脸动画是计算机图形学、计算机视觉领域最根本、最困难的问题之一,也是近年来国际上最活跃的研究热点。真实感人脸建模的研究在电影、广告、人物动画、计算机游戏、视频会议、人机交互、医学、公安安全等许多领域有广阔的应用前景。形变模型(Morphable Model)是近几年出现的三维人脸建模新方法。虽然该模型同以往的参数模型、肌肉模型以及视觉模型相比,具有真实感好、自动化等特点,但目前该模型在模型建立、优化以及光照等方面还存在诸多问题,在模型应用方面也缺乏探索。针对上述问题,本文在基于形变模型的人脸建模方法及应用方面展开了研究,主要的研究内容包括: (1) 面向中国人特征的大规模三维人脸数据库的建立。三维人脸数据库是建立形变模型的基础。数据库的建立包括数据获取、数据预处理、数据规格化以及数据库结构设计等研究内容。关键是要解决高维三维人脸数据基于特征的像素级稠密对应问题。 (2) 基于形变模型的人脸建模方法及改进。首先,基于三维人脸数据库,研究三维人脸形变模型的建立方法,主要包括参数模型的建立和针对输入人脸图像的模型匹配。模型建立涉及模型基础数据的量化处理和人脸参数表示模型的建立,模型匹配涉及光照模型、摄像机标定以及大规模优化问题。其次,对现有的形变模型进行了改进,主要针对非理想光照环境下的模型匹配问题和匹配优化加速问题展开研究。 (3) 形变模型的应用研究。主要研究模型在人脸动画、人脸识别和人脸特征提取方面的应用。在人脸动画方面,研究人脸动画的参数表示模型和人脸动画的驱动方法。在人脸识别方面,针对人脸识别中的光照和姿态问题,研究基于三维人脸模型的人脸识别方法。在人脸特征提取方面,研究形变模型在二维人脸图像特征表示方面的应用。
[Abstract]:Realistic 3D face modeling and face animation is one of the most fundamental and difficult problems in the field of computer graphics. It is also the most active research hotspot in the world in recent years. The research of real moving face modeling has broad application prospect in many fields such as film, advertisement, character animation, computer game, video conference, human-computer interaction, medicine, public security and so on. Deformation Model is a new method of 3D face modeling in recent years. Compared with the previous parametric model, muscle model and visual model, the model has the characteristics of reality and automation, but there are still many problems in the modeling, optimization and illumination. There is also a lack of exploration in the application of the model. In order to solve the above problems, the method and application of face modeling based on deformation model are studied in this paper. The main research contents include: 1) the establishment of a large scale 3D face database for Chinese features. 3D face database is the foundation of building deformation model. The establishment of database includes data acquisition, data preprocessing, data normalization and database structure design. The key is to solve the feature-based pixel level dense correspondence problem of high dimensional 3D face data. (2) face modeling method based on deformation model and its improvement. Firstly, based on the 3D face database, the method of establishing 3D face deformation model is studied, including the establishment of parameter model and the model matching for input face image. Modeling involves the quantization of the basic data of the model and the modeling of facial parameter representation. Model matching involves illumination model, camera calibration and large-scale optimization. Secondly, the existing deformation model is improved. This paper focuses on the model matching problem and the matching optimization acceleration problem in the non-ideal illumination environment. This paper mainly studies the application of the model in face animation, face recognition and face feature extraction. In the aspect of face animation, the parameter representation model of face animation and the driving method of face animation are studied. In the aspect of face recognition, aiming at the illumination and pose problem in face recognition, a face recognition method based on 3D face model is studied. In the aspect of face feature extraction, the application of deformation model in 2D face image feature representation is studied.
【学位授予单位】：北京工业大学
【学位级别】：博士
【学位授予年份】：2004
【分类号】：TP391.4

【引证文献】