I want to use my own dataset/poses without running COLMAP. I have a few questions, apologies because I am a little new to this:
For generating a poses_bounds.npy file, what does it mean that "we assume the principal point is centered"? What are the transformation matrices relative to for each image?
I assume all images need to be perfectly centered on the object being pictured?
What are the dimensions for the "close/far depth bounds" inputs?
Hi,
I want to use my own dataset/poses without running COLMAP. I have a few questions, apologies because I am a little new to this: