Open gigasurgeon opened 2 months ago
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?
Hey,
You can accelerate it a lot by setting fine_tracking=False, which should be not necessary for video input
fine_tracking=False
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?