Efficient method for face region quality enhancement in low bit rate video coding

This paper presents an efficient method for the enhancement of face region image quality. The benefit of it in complex computation, preserving necessary background region information and no blocking effect make this method outperform the current existing method. Furthermore, this method provides adaptive weighting factors and a multi background region approach that allows one to control the effect of the filter as a function of available channel bit rate. An improved face region segmentation method is also addressed in this paper. Simulation results confirmed that the proposed method has significantly improved image quality of face region (up to 2.65 dB) and reduced a large number of skipped frames in comparison with ITU-T H.263+ video coding test model 10 (TMN10).