Is this method suitable for 3DScene? example, a room
zuixiaosanlang opened this issue · comments
I try, but get a bad result.
Hi,
We have not tried it, but I do not think it can work as well as our object-level reconstruction results. One main reason is the limited resolution of the indicator grid. Another reason is, for the learning-based pipeline, we only trained on the ShapeNet object, not indoor scene dataset like this synthetic indoor scene dataset. If you can train on that, I believe you can have better results.
Best
Songyou