Video Painting based on a Stabilized 3D Flow Field

We present a method for generating a 3D feature flow field from a video and its application to video stylization. Our method extracts smoothly aligned 3D vectors that describe the smallest variation of colors in both spatial and temporal dimensions of the video, and thus efficiently preserves both spatial and temporal coherence in a relatively inexpensive manner. We use this flow field forms the basis of a particle-based video stylization technique which can produce feature-enhancing painting-style renderings of a video. Furthermore, we show our method is performed in real-time using the GPU based implementation.