Recovering 3D color and motion information from multi-view 2D images

• Main Author: Pham, Xuan Hoa.
• Other Authors: Cai Jianfei
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/38567
• Institution: Nanyang Technological University

Recovering 3D color and motion information of human is a fundamental problem in many applications, including realistic character animation for games and movies, motion analysis for medical diagnostics and sport scene, reconstruction of human appearance and motion for 3D virtual reality. The first step of solving this problem is to recover human information from a sequence of 2D images captured by cameras from many viewpoints. There are many ways to recover color information from multi-view images. One way is to store color information in a texture image, which needs associated texture coordinates of 3D model vertices. Therefore, a parameterization method is necessary. The ABF++ method in Graphite software is directly applied due to the complexity of parameterization. After that, all color information from multi-view images is retrieved and blended together to create a texture image.