Skip to content
This repository has been archived by the owner on Nov 3, 2021. It is now read-only.
/ M3-LRN Public archive

Accurate 3D Facial Geometry Prediction by Multi-Task, Multi-Modal, and Multi-Representation Landmark Refinement Network

License

Notifications You must be signed in to change notification settings

choyingw/M3-LRN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

e7c038b · Oct 20, 2021

History

13 Commits
Apr 20, 2021
Apr 16, 2021
Apr 16, 2021
Oct 20, 2021

Repository files navigation

M3-LRN

Arxiv 2021: Accurate 3D Facial Geometry Prediction by Multi-Task, Multi-Modal, and Multi-Representation Landmark Refinement Network

Cho-Ying Wu, Qiangeng Xu, Ulrich Neumann, CGIT Lab at University of Souther California

[paper] [project page]

Advantages:

+ State-of-the-art on all 3D facial alignment, face orientation estimation, and 3D face modeling.

+ Exploitation of Multi-Modal and Multi-Representation for information aggregation.

+ Fast and easy to use: 3000fps for 3D facial landmarks on a single GPU.

+ The first to study the face reconstrcutability from sparse landmarks.

More results:

Facial alignemnt on AFLW2000-3D (NME of facial landmarks):

Face orientation estimation on AFLW2000-3D (MAE of Euler angles):

General results:

Updates: The codes are now released https://github.com/choyingw/SynergyNet. This repo will be archived due to its older version.

About

Accurate 3D Facial Geometry Prediction by Multi-Task, Multi-Modal, and Multi-Representation Landmark Refinement Network

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published