Scanned depth

[dnlwbr]

In your video you said, that you tested DSNeRF with scanned depth data. How can I train the model with my own depth data. Which format is needed? Thanks in advance.

[bigdimboom]

The input in the paper is sparse 3d/depth. I think they are extending the concept to using dense depths in the video? I guess the challenge is how you gonna acquire the weights for the dense depths,. You would still need compute reprojection error for them. I think even the SOA depth sensors have certain amount of errors.

[dnlwbr]

Wouldn't it be possible to just use a constant weight for all depth points? I think the Redwood dataset used in the paper does not provide confidence/error values either.

Unfortunately, the code in this repo seems to be incomplete and different from the code used in the paper, since the function load_sensor_depth() in load_llff.py is almost the same as load_colmap_depth().

[mertkiray]

Any update on how to use sensor depth?

Also, I am using ndc space and I have depth values collected from a sensor, should I normalize these values between 0 and 1 or should I convert them into ndc space?

[endlesswho]

I just want to test whether blender lego depth file works for this frameworks, only 2 view for better novel view generation!

[endlesswho]

@dunbar12138

I just want to test whether blender lego depth file works for this frameworks, only 2 view for better novel view generation!

[MaxChanger]

Hi guys, is there any progress on raw seneor depth as input An immature idea is to convert the pose and depth image into the colmap bin file format by a script, but I think this is superfluous, rewriting load_sensor_depth may be better, but the changes involved may be more