search

vevenom / RoomLayout3D_RandC

https://www.tugraz.at/index.php?id=40222
BSD 3-Clause Clear License
48 stars 12 forks source link

Depth image format #8

Open dilpreet7365 opened 1 year ago

dilpreet7365 commented 1 year ago
  1. What should be the preprocessing step for the depth images ? The values within the depth images are really low how do we get the images in the same format.
  2. How can we run the model with out passing the depth image as an input
vevenom commented 1 year ago
  1. I am not sure I understand your question.
  2. You can use depth prediction networks to predict the depth. You can find more details in the paper.
dilpreet7365 commented 1 year ago

Thank you for your reply.

I was using the depth predicted by planeRCNN as input but I think there are some preprocessing steps involved in the depth image format because, the values of the depth are lying between 1-20 in your example dataset. Can you help me with the preprocessing step? What should be the depth image format for the input.

Also, can we generate 3D layout without passing depth image as an input?

Regards, Dilpreet Kaur

On Tue, 14 Feb 2023 at 5:57 PM, vevenom @.***> wrote:

  1. I am not sure I understand your question.
  2. You can use depth prediction networks to predict the depth. You can find more details in the paper.

— Reply to this email directly, view it on GitHub https://github.com/vevenom/RoomLayout3D_RandC/issues/8#issuecomment-1429669564, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHG3OBHPLG5CYJBGXFZMGQTWXN225ANCNFSM6AAAAAATXGYYYQ . You are receiving this because you authored the thread.Message ID: @.***>