search

AaronJackson / vrn

:man: Code for "Large Pose 3D Face Reconstruction from a Single Image via Direct Volumetric CNN Regression"
http://aaronsplace.co.uk/papers/jackson2017recon/
MIT License
4.52k stars 747 forks source link

Results in vrn-docker #159

Closed seven-sent closed 3 years ago

seven-sent commented 3 years ago

Hello, thank you so much for sharing your wonderful work. I am new in 3D face reconstruction and I have some doubts when using vrn-docker.

I successfully ran the vrn-docker and got the result obj files, but the results did not look good. When I open the obj files in meshlab, I got this error:

Error encountered while loading file: "001.jpg.obj" Error details: Identical vertex indices found in the same faces -- faces ignored

I ignored the error and open the file: Snipaste_2021-07-04_15-58-43

This seems strange. Is this the right result, or what caused it? How can I get better results?

AaronJackson commented 3 years ago

There's a bit of noise, happens sometimes. The face looks blocky because the marching cubes implementation available in Python doesn't make use of non-binary values. Matlab has a nicer function for doing this