search

3dlg-hcvc / paris

[ICCV 2023] Official implementation of the paper "PARIS: Part-level Reconstruction and Motion Analysis for Articulated Objects"
https://3dlg-hcvc.github.io/paris/
MIT License
61 stars 2 forks source link

scale between start and final #2

Closed Guptajakala closed 1 year ago

Guptajakala commented 1 year ago

Hi, very cool work!

It's said only RGB is needed. How are the scales ensured to be consistent between the scans for the start and end states?

SevenLJY commented 1 year ago

Hi! Thanks for your interest in our work!

As you might notice, it is an assumption in our problem setting that the objects in the start and end states have been aligned in the world coordinate system. For the experiments shown in the paper, the scale consistency comes with the datasets. In a real scanning setting, this can be ensured by sharing at least one input view across two states. That being said, as long as the camera is fixed in the sharing view and the object is fixed (and only articulating the part), then the scale is consistent.

Hope it helps!