Is there a way to make the camera path estimation faster?

[gigasurgeon]

I am working with a use case where only camera path is required in a video. I do not want the sparse reconstruction. Is there a way to skip reconstruction to make the pipeline faster?

[jytime]

Hey,

You can accelerate it a lot by setting fine_tracking=False, which should be not necessary for video input