[QUESTION] about MST Initialization in Global Aligner

[yukiumi13]

Thank you for your excellent work on this project.

I am having some confusion regarding the initialization process in the following lines of code:

dust3r/dust3r/cloud_opt/init_im_poses.py

Lines 150 to 155 in 69192aa

	while todo:
	# each time, predict the next one
	score, i, j = todo.pop()
	if im_focals[i] is None:
	im_focals[i] = estimate_focal(pred_i[i_j])

It appears that a new edge (i, j) is popped, but the variable i_j is not updated accordingly. Additionally, it seems that the estimated focal length of the legacy i is being assigned to the newly popped i view. Could you please clarify if this is the intended behavior or if there might be an issue here?

I would appreciate any insights you can provide to help me understand this better.

Thank you!