TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
3D Face Reconstruction	Florence	Genova et al.	RMSE Cooperative	1.78	# 7
3D Face Reconstruction	Florence	Genova et al.	RMSE Indoor	1.78	# 7
3D Face Reconstruction	Florence	Genova et al.	RMSE Outdoor	1.76	# 4
3D Face Reconstruction	Florence	Unsupervised-3DMMR	Average 3D Error	1.50	# 2

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/unsupervised-training-for-3d-morphable-model/3d-face-reconstruction-on-florence)](https://paperswithcode.com/sota/3d-face-reconstruction-on-florence?p=unsupervised-training-for-3d-morphable-model)`

Unsupervised Training for 3D Morphable Model Regression

CVPR 2018 · Kyle Genova, Forrester Cole, Aaron Maschinot, Aaron Sarna, Daniel Vlasic, William T. Freeman ·

We present a method for training a regression network from image pixels to 3D morphable model coordinates using only unlabeled photographs. The training loss is based on features from a facial recognition network, computed on-the-fly by rendering the predicted faces with a differentiable renderer. To make training from features feasible and avoid network fooling effects, we introduce three objectives: a batch distribution loss that encourages the output distribution to match the distribution of the morphable model, a loopback loss that ensures the network can correctly reinterpret its own output, and a multi-view identity loss that compares the features of the predicted 3D face and the input photograph from multiple viewing angles. We train a regression network using these objectives, a set of unlabeled photographs, and the morphable model itself, and demonstrate state-of-the-art results.