search

autonomousvision / convolutional_occupancy_networks

[ECCV'20] Convolutional Occupancy Networks
https://pengsongyou.github.io/conv_onet
MIT License
835 stars 114 forks source link

Different input number of points for reconstruction task #3

Closed Wuziyi616 closed 4 years ago

Wuziyi616 commented 4 years ago

Hi! Thanks for your great work! Recently I want to apply your method to 3D reconstruction from very sparse input point clouds (often less than 1000 points). And I discover that you use 3000 points in "configs/pointcloud/shapenet_3plane.yaml". Actually I've tried ONet with 300 input points before and its performance in my task is okay. So I wonder have you tried ConvONet with fewer input points, and what's its IoU result?

I understand that because you need to project point features to grids and then perform convolution, so more points would be better. I just wonder if you've tried other input point number in your experiments and whether its performance is sensitive to this. Thank you!

pengsongyou commented 4 years ago

Hi Ziyi,

Thanks for your interest in our work. I think I tried at the early stage of this project with fewer points, like 300, and if I remember correctly, our quantitative results were better than ONet but not much. This makes sense because sparse point clouds do not give enough cues to the objects' geometry. However, the training speed of our model is definitely much faster than ONet in any case. I encourage you to try it out with sparse point clouds yourself.

Wuziyi616 commented 4 years ago

Yes, thank you for your quick reply. Will try it later by myself.