CVPR 2023

High-Res Facial Appearance Capture from Polarized Smartphone Images

1Technical University of Munich    2Meta Reality Labs    3Max Planck Institute for Intelligent Systems

Our method obtains high-resolution skin textures from two RGB input sequences captured with polarization foils attached to a smartphone.


We propose a novel method for high-quality facial texture reconstruction from RGB images using a novel capturing routine based on a single smartphone which we equip with an inexpensive polarization foil. Specifically, we turn the flashlight into a polarized light source and add a polarization filter on top of the camera. Leveraging this setup, we capture the face of a subject with cross-polarized and parallel-polarized light. For each subject, we record two short sequences in a dark environment under flash illumination with different light polarization using the modified smartphone. Based on these observations, we reconstruct an explicit surface mesh of the face using structure from motion. We then exploit the camera and light co-location within a differentiable renderer to optimize the facial textures using an analysis-by-synthesis approach. Our method optimizes for high-resolution normal textures, diffuse albedo, and specular albedo using a coarse-to-fine optimization scheme. We show that the optimized textures can be used in a standard rendering pipeline to synthesize high-quality photo-realistic 3D digital humans in novel environments.



We use our reconstructed geometry and textures to render a face from novel views and under novel illumination. The recovered assets are ready to be used in off-the-shelf rendering software, such as Blender, for photo-realistic relighting.

We visualized the captured faces from novel views and with novel lighting. In this example, the faces are lit by a point light that is moving together with the camera.
We relight the faces with a point light that is moving independently of the camera. This is a novel scenario since the optimizer had access only to frames where the camera and light were co-located.
We import the reconstructed geometry and textures into Blender. The faces are then re-lit by an HDR environment map (visualized in the bottom-left corner).


    author    = {Azinovi\'c, Dejan and Maury, Olivier and Hery, Christophe and Nie{\ss}ner, Matthias and Thies, Justus},
    title     = {High-Res Facial Appearance Capture from Polarized Smartphone Images},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2023}