3D Face Reconstruction System Based on Deep Learning and Sparse Face Model
Download as PDF
DOI: 10.23977/acsat.2017.1011
Author(s)
Sun Guangmin, Shi Xiuwen, Sun Yuge
Corresponding Author
Guangmin Sun
ABSTRACT
3D face reconstruction technology is very popular in the digital image processing area. The method of 3D face reconstruction based on single image faces many challenges such as: (i) depth information is lacked due to the input of 2D image; (ii) the statistical face model is not accurate; and (iii) many current difficulties can be solved by using new technique like deep learning. In order to get an accurate and efficient 3D face reconstruction result, a new 3D face reconstruction algorithm based on deep learning and sparse 3D face model is proposed in this paper. Deep learning is exploited to find out the statistical property of 3D human face. And sparse 3D face model is applied to improve algorithm efficiency. Experiments under different conditions are performed to prove the accuracy and robustness of our proposed algorithm.
KEYWORDS
Face Reconstruction, Deep Learning, Sparse 3D Face Model.