We have a doubt about the evaluation of the foreground detection after video stabilization. If we use the given groundtruth it won't match the stabilizated video, so it does not make sense to evaluate it.
Should we also apply the stabilization algorithm to the groundtruth?
Yes, in order to make a proper evaluation you have to either stabilized the ground truth or de-stabilized the result before evaluate it. Make sure you explain your selection in your report!
We have a doubt about the evaluation of the foreground detection after video stabilization. If we use the given groundtruth it won't match the stabilizated video, so it does not make sense to evaluate it. Should we also apply the stabilization algorithm to the groundtruth?