Video Stabilization
This algorithm is a patch based stabilization and a refactored version of the code from [1]. Finding the corresponding matches for compensating shaking goes patch by patch, which means that each patch has surely one feature. It is the different point with other algorithms whose feature points are not evenly distributed. I guess this is one of applications from [2].
References
[1] https://github.com/LMescheder/Video-Stabilizer
[2] Richard Szeliski. 2010. Computer Vision: Algorithms and Applications (1st ed.). Springer-Verlag, Berlin, Heidelberg
