3D face reconstruction of shape and skin texture from a single 2D image can be performed using a 3D Morphable Model (3DMM) in an analysis-by-synthesis approach. However, performing this reconstruction (fitting) efficiently and accurately in a general imaging scenario is a challenge. Such a scenario would involve a perspective camera to describe the geometric projection from 3D to 2D, and the Phong model to characterise illumination. Under these imaging assumptions the reconstruction problem is nonlinear and, consequently, computationally very demanding. In this work, we present an efficient stepwise 3DMM-to-2D image-fitting procedure, which sequentially optimises the pose, shape, light direction, light strength and skin texture parameters in separate steps. By linearising each step of the fitting process we derive closed-form solutions for the recovery of the respective parameters, leading to efficient fitting. The proposed optimisation process involves all the pixels of the input image, rather than randomly selected subsets, which enhances the accuracy of the fitting. It is referred to as Efficient Stepwise Optimisation (ESO).
This work was published in Pattern Recognition.
- Full-text PDF (3.43 Mb) -- cite as:
G Hu, F Yan, J Kittler, W Christmas, C Chan, Z Feng, P Huber, Efficient 3D Morphable Face Model Fitting, Pattern Recognition, 67:366-367, 2017