Will your theory be inaccurate on the edge

patrikhuber / eos

A lightweight 3D Morphable Face Model library in modern C++

Apache License 2.0

1.89k stars 596 forks source link

Will your theory be inaccurate on the edge #305

Closed sunjunlishi closed 4 years ago

sunjunlishi commented 4 years ago

Your algorithm is refined and the code is simple.

Using 3D rotation translation and other operations, project to 2D, and then correct with 2D key points.

The 17 points of the starting 3D edge, after rotation, will go around to the back, and the key points of 2D will not coincide, which is the problem. Does your SFM _edge_topology.json solve this problem? QQ图片20200419123254

patrikhuber commented 4 years ago

Hi @sunjunlishi,

You can find the details of the contour fitting in my PhD thesis, there's a section on contour fitting. Additionally, the code should be fairly well documented and it's based on the paper by Bas et al. (see the eos README.md). The files in the share/ directory are also further explained by readme.txt in that same directory.

Thanks.

sunjunlishi commented 3 years ago

thank you very much,author