MPEG-4 compression artifacts removal on color video sequences using 3D nonlinear diffusion

Today's needs for low bitrates in data compression are responsible for the MPEG and ITU family of video coding standards' success. However, it is important to realize that such standards involve lossy compression. This means a reconstructed image sequence will contain degradations relative to the original, because information is discarded during compression. Lower bitrate compression modes often introduce block edge, ringing, blurring, color bleeding, or mosquito noise effects. In this paper, we propose a partial differential equations-based (PDE) method to reduce compression-related artifacts on MPEG-4 color image sequences. The main advantage of our approach is that both temporal and spatial artifacts removal are performed at the same time, by considering an image sequence as a 3D object. Although PDE-based methods for still images restoration are becoming quite popular, extensions to image sequences remain rare, especially in color. This is also the purpose of this paper to introduce one.