Awesome

Purpose

this is a Matlab implementation of Functional Faces: Groupwise Dense Correspondence using Functional Maps by Chao Zhang, William A.P. Smith, etc

Dependency

the code is developed based on some existing codes, libraries, and data, they include:

1. pairwise func. map implementation developed by M. Ovsjanikov as to the paper Functional maps: a flexible representation of maps between shapes.
2. manopt: a Matlab toolbox for optimization on manifolds
3. the sample code uses three meshes from a high quality facial dataset: reference: G. Stratou, A. Ghosh, P. Debevec, and L. Morency. Effect of illumination on automatic expression recognition: a novel 3D relightable facial database. In Proc. Face and Gesture, pages 611–618, 2011

Usage

1. Download manopt if you were not using it and set up path in the code (see demo_ff.m)
2. Choose LB dimension and feature functions (see getMapOfTwoFaceShapes.m)
3. Run demo_ff.m

Note

1. to compute the point-wise correspondence from functional maps, approximate nearest neighbor search might need to be compiled per to the operating system

Reference

@inproceedings{zhang2016functional, title={Functional Faces: Groupwise Dense Correspondence using Functional Maps}, author={Zhang, Chao and Smith, William AP and Dessein, Arnaud and Pears, Nick and Dai, Hang}, booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition}, pages={5033--5041}, year={2016} }