FLAME
- Statistical face model with PCA basis for shape, expression and pose, and jaw and neck joints
- 300-dimensional shape basis from 3800 scans
- Pose is modeled using 9-dimensional pose basis and modified using linear blend skinning
3D Face Reconstruction from Dense Landmarks
- Energy based method for face reconstruction with 703 landmarks trained on synthetic data
- First half of the pipeline estimates 2D landmarks, second half of the pipeline fits the 3D model
- Fitting of 3D model requires no image supervision
3D Face Reconstruction with Weakly Supervised Learning
- This work leverages a hybrid loss function for weakly-supervised learning which takes into account both low-level and perception-level information for supervision
- Does not require 3D facial data for training as the loss is calculated between the rendered 2D Image from the model’s 3D Face and the input 2D Image
References
Li, Tianye et al. “Learning a model of facial shape and expression from 4D scans.” ACM Transactions on Graphics (TOG) 36 (2017): 1 – 17.
Wood, Erroll, et al. “3d face reconstruction with dense landmarks.” European Conference on Computer Vision. Cham: Springer Nature Switzerland, 2022.
Deng, Yu, et al. “Accurate 3d face reconstruction with weakly-supervised learning: From single image to image set.” Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops. 2019.