how is pose estimation generated ?

[an99990]

Hi,
As I am reading your paper, I dont seem to be able to find how you guys estimate the pose of the robot after your removed the outliers.
What happens next ? in the pipeline it says visual odometry ? and then octo-map generation.

Can you guys explain how you generate the pose ?