Jing Zeng

How to improve image quality by monosdf rendering

7

have a multi-room scenario, using 400 images for reconstruction with MonoSDF. The rendered new viewpoints only achieve a PSNR of 21. How can I improve this?  

The exploration trajectory encounters a collision

4

Thanks for your nice work， When I run ii in scene: office.pcd, the exploration trajectory encounters a collision. Why does this situation occur, and how can we resolve this issue？...

No instant_pred results for scans_test split (not val and without label) dataset ？

1

When I test scans_test split dataset (not val and without label) dataset, `python visualization.py --prediction_path ../results/ --room_name scene0707_00 --task instant_pred --out ../results/scene0707_00_instant_pred.ply` output is  but `python visualization.py --prediction_path ../results/...

Low training speed (0.5 fps)

10

When I run it in room0 scene， speed is about 0.5 fps， which is much lower than the paper， why is it? 

Image quality by rendering

3

I have a problem, Image quality by rendering VS monosdf and other nerf based method ? In replica room0 dataset , is it possible to achieve PSNR > 30 such...

How to compute the scale of omnidata depth?

15

Thanks for your work, but I do not find the code to compute the scale of omnidata depth in the paper. 

Does it support outdoor scenes?

1

Does it support indoor scenes? Or are there limitations such as network forgetting issues when using larger multi-room scenes like Matterport 3D dataset as below 

ModuleNotFoundError: No module named 'prim3d'

1

using scene_id: 0616_00 train_suffix: 0616_00_pretrained vis_suffix: 0616_00_pretrained mvs_suffix: mvs **kwargs: --casting Traceback (most recent call last): File "scripts/train_pretrained.py", line 17, in from scripts.generate_normal import ray_casting_depth_normal File "/mnt/dataset/zengjing/HelixSurf/scripts/generate_normal.py", line 9, in...