Stabbo

Excellent work. But, wouldn´t you release the source ? It would be good others could help you improving the app. There are some issues that needs to fix, like, the background of some scenes are moving improperly when they should be fixed. Also, there is no need to zoom the image since you can collect the data surround the borders to fill the blanks. One thing that may be needed to help the development is implement a scene detector from where the user can manage how the stabilization will be done on each scene. Example, let´s say that a video (short clip) have 20 scenes. Each one of them are moving separately from each other. Scene 1 the background can be moving forward but the central objects (a person) may be standing still. On scene 2 it can happens the opposite, the background is fixed (not moving) but the person is moving. On scene 3 the background is not moving, but entered on the scene another element (other person) etc etc. Scene detection can be a good way to make the app finds as many objects (elements) as possible on each scene and upon then the user can choose what to do with them. I mean, let´s say that on Scene 1 the app detected 5 different objects plus the background, then you can make the program isolate each object to the user do what he wants with them. For example, he can remove one of the objects from the whole scene or remove them all and let only the background etc. Also, object detection can also be used to detect and remove logos.