Yifan Lu

I don't think (img.camera_readout_done_time - img.camera_trigger_time) is the exposure time for each pixel. There is another attribute img.shutter.

@renxiangdai @tsattler Hi, I am also curious about this problem. Do we also have the 3D point - to - image correspondence in the dense reconstruction stage? I want to...

I try to replace the `/colmap/dense/images` with semantic maps before the `stereo_fusion` step. It gets some results that look right, but the color of 3D points will not strictly abide...

I fix it. It is due to the .jpg compression.

Other method like nerfacto can work well with my dataset with `--auto-scale-poses False`.

@maturk Hi! I think I have put `colmap` before `--auto-scale-poses False`. How should I modify my command? Thanks!

Hi! Do you find out a solution to this?

You may have a bad initialization, no points fall in the camera views. @qianyuling-wl

@cs-mshah `FovX` and `FovY` do not need rescale because they are always there. But when you use the intrinsics `K`, you need to rescale it according to your new resolution.