Human face fast-3D-reconstruction from a single image

摘要：

Human face three-dimensional (3D) reconstruction is a challenging problem. In this paper, we propose a human face fast- 3D- reconstruction method based on image processing with a single image. Shape from shading (SFS) is chosen to reconstruct the human face. First, SFS theory is introduced. It has the advantage of fast 3D reconstruction and only need a single image. Secondly, because the noise will affect the 3D reconstruction result greatly, wavelet transform and wavelet packet transform are introduced and used in image denoising respectively. The experiment has shown that the method based on wavelet transform produces the best denoising result than wavelet packet transform. At last, a human face 3D reconstruction algorithm based on a single image is proposed. The experimental results show that a human face 3D model can be reconstructed in fast by proposed algorithm.

关键词： Human face Three-dimensional reconstruction Shape from shading Wavelet packet transform

作者: Huihuang Zhao Yaonan Wang Yaqi Sun Jianzhen Chen

作者单位: Department of Computer, Hengyang Normal University, HuNan, HengYang, 421008, China Department of Con Department of Control Science and Engineering, Hunan University, Hunan 410082, China Department of Computer, Hengyang Normal University, HuNan, HengYang, 421008, China

会议类型: 国际会议

会议名称: 2011 International Conference on Mechatronics and Applied Mechanics(2011年机电一体化与应用力学国际会议 ICMAM 2011)

会议地点: 香港

会议语种:英文

页码: 1008-1011

在线出版日期: 2011-12-27（万方平台首次上网日期，不代表论文的发表时间）