Creating animations combining stochastic paintbrush transformation and motion detection

In this paper we propose a method for creating animation and animation-like video sequences from any ordinary video recorded by any means available. The method is based on paintbrush transformation, developed and patented by our department, and on different motion detection algorithms. One of the goals of the method is to obtain animations, cartoon-like outputs from a usual camera-recorded image sequence. The method inherits properties of the paintbrush transformation method like well-defined contours, acceptable distortion, and a painting-like view with no fine details below a limit. The resulting output is a series of video frames stored as brush-strokes and motion data between the frames, compressed for size reduction. This output can be converted to any available video format by decompression, frame-reconstruction and recompression.