What did you do with the dense depth map in the result

zzangjinsun / NLSPN_ECCV20

Park et al., Non-Local Spatial Propagation Network for Depth Completion, ECCV, 2020

MIT License

321 stars 55 forks source link

What did you do with the dense depth map in the result #36

Closed zura-false closed 3 years ago

zura-false commented 3 years ago

Hello, it's a great honor to refer to your work，I'm making subsequent point cloud processing for your depth completion results，When using the pictures posted in the paper, I can succeed，However, when I use the results you published in result or the results of my own test, I will fail。So I want to know what you did with the depth completion results，thank you。

zzangjinsun commented 3 years ago

Sorry I didn't get the point of your question.

I think you may have different input value ranges between images [0, 255] and depth completion results [0, 80~100]?