search

michaelStettler / FacialRetargeting

Automatic Facial Retargeting
MIT License
61 stars 8 forks source link

can you show some result？ #2

Open ZHANG-SHI-CHANG opened 3 years ago

ZHANG-SHI-CHANG commented 3 years ago

good job！ can you show some result of your job?

ZHANG-SHI-CHANG commented 3 years ago

good job！ can you show some result of your job?

by the way, i find 4.4, 4.5 and 5 chapter have wrong programm, maybe you have wrong understand about laplacian

michaelStettler commented 3 years ago

Hi,

Sorry as I answered to the other open issues, I am not able to show much since we have not published anything yet.

About the Laplacian, may I ask if you have any test case I could try on? I remember implementing the computation of the Lapalican based on the work mentioned in the script, which I have then tested it against the pymesh library.

ZHANG-SHI-CHANG commented 3 years ago

Hi,

Sorry as I answered to the other open issues, I am not able to show much since we have not published anything yet.

About the Laplacian, may I ask if you have any test case I could try on? I remember implementing the computation of the Lapalican based on the work mentioned in the script, which I have then tested it against the pymesh library.

hi, I mean laplacian in 4.4, 4.5 and 5 chapter loss should use dot product not multiply

michaelStettler commented 3 years ago

Hi,

I have to look then. But the dot product is simply the sum of the multiplication, are you sure that I am not applying the sum beforehand or latter on using the repetitions tricks? If you have some specific lines you think are faulty it would be faster for me to have a look :)

ZHANG-SHI-CHANG commented 3 years ago

Hi,

I have to look then. But the dot product is simply the sum of the multiplication, are you sure that I am not applying the sum beforehand or latter on using the repetitions tricks? If you have some specific lines you think are faulty it would be faster for me to have a look :)

hi， for example，Equation 11 and Equation 13 consider the relationship of vertices and blendshape, multiplication can not show this，and the relationship is very importance for laplacian