Hello, I am testing refusion on OpenLORIS dataset, which is an indoor RGBD dataset collected via realsense D435i.
And it's obvious that refusion performs very bad when the camera is turning. And without color, it performs very bad.
I wonder whether this method really can't estimate pose well?
Hello, I am testing refusion on OpenLORIS dataset, which is an indoor RGBD dataset collected via realsense D435i. And it's obvious that refusion performs very bad when the camera is turning. And without color, it performs very bad. I wonder whether this method really can't estimate pose well?
Thank you very much! Best regards