How to export shape/AO/depth image from trained instant-ngp

[yellowYuga]

Thank you for this great work. However, I encountered a problem while building my own dataset. I can use your code to generate a new transforms.json file and train Instant NGP successfully. But I'm wondering how to export RGB and AO/depth images from the trained Instant NGP. Do I need to use transforms.json and transforms_left.json separately to generate different training results for nerf? By the way, I am using the GUI on Windows.

[fabiotosi92]

Hi, in our project, we obtained all the necessary information from Instant-NGP by directly using the command line instead of relying on the GUI. Specifically, we utilized the code provided in scripts/run.py. Within this code, you have the flexibility to modify the render_mode argument based on your requirements. You can choose between testbed.render_mode = ngp.Depth, testbed.render_mode = ngp.AO, or leave it as the default to render RGB images.

However, it is important to note that the updated version of the Instant-NGP code might have undergone changes since we used it for data generation.

PS. Yes, we used transforms_left.json and transforms.json separately for the data generation.

[yellowYuga]

Thank you, I have been able to render based on the JSON file you provided. Although the rendered images have some artifacts and appear slightly blurry, I will examine the input images of instant-ngp.

[Qjizhi]

Hi, in our project, we obtained all the necessary information from Instant-NGP by directly using the command line instead of relying on the GUI. Specifically, we utilized the code provided in scripts/run.py. Within this code, you have the flexibility to modify the render_mode argument based on your requirements. You can choose between testbed.render_mode = ngp.Depth, testbed.render_mode = ngp.AO, or leave it as the default to render RGB images.

However, it is important to note that the updated version of the Instant-NGP code might have undergone changes since we used it for data generation.

PS. Yes, we used transforms_left.json and transforms.json separately for the data generation.

Hi Fabio,

First thanks for your excellent work! I can get the left and right images now. However, the depth map are white and black if I only use testbed.render_mode = ngp.Depth and get this depth map if I use

testbed.render_mode = ngp.Depth
testbed.exposure = -3.8

The value ranges [0, 255], do you know how to get absolute depth?

[CaptainEven]

Hi, in our project, we obtained all the necessary information from Instant-NGP by directly using the command line instead of relying on the GUI. Specifically, we utilized the code provided in scripts/run.py. Within this code, you have the flexibility to modify the render_mode argument based on your requirements. You can choose between testbed.render_mode = ngp.Depth, testbed.render_mode = ngp.AO, or leave it as the default to render RGB images. However, it is important to note that the updated version of the Instant-NGP code might have undergone changes since we used it for data generation. PS. Yes, we used transforms_left.json and transforms.json separately for the data generation.

Hi Fabio,

First thanks for your excellent work! I can get the left and right images now. However, the depth map are white and black if I only use testbed.render_mode = ngp.Depth and get this depth map if I use

testbed.render_mode = ngp.Depth
testbed.exposure = -3.8

The value ranges [0, 255], do you know how to get absolute depth?

Have you solved the problem yet?

[zhaoxin111]

Hi, in our project, we obtained all the necessary information from Instant-NGP by directly using the command line instead of relying on the GUI. Specifically, we utilized the code provided in scripts/run.py. Within this code, you have the flexibility to modify the render_mode argument based on your requirements. You can choose between testbed.render_mode = ngp.Depth, testbed.render_mode = ngp.AO, or leave it as the default to render RGB images.

However, it is important to note that the updated version of the Instant-NGP code might have undergone changes since we used it for data generation.

PS. Yes, we used transforms_left.json and transforms.json separately for the data generation.

Are you referring to the scripts/run.py in instant-ngp? There is no render_mode parameter that can be set in the current code.