Open ghost opened 2 years ago
I haven't tested my code on shapenet. But you can have a try! What you need for inversion:
Hi, I'm trying to get the inverse code for a vehicle image, curious if this project supports inverse for shapenet datasets?
I managed to get quite nice results for the w and w+ from Shapenet, but I am still working on the PTI inversion and generating the final video. After you get the camera parameters and testing image, you have to change the generator resolution and some other geometry related stuff (eg. focal length) which is kind of hard-coded in this code.
Hi, I'm trying to get the inverse code for a vehicle image, curious if this project supports inverse for shapenet datasets?