Adaptive Spatio-Temporal Filtering for Video De-noising (MP-P5)

Author(s) : Hye-Yeon Cheong (Columbia University, USA) Alexis Michael Tourapis (Thomson Corporate Research, USA) Joan Llach (Thomson Corporate Research, USA) Jill Boyce (Thomson Corporate Research, USA)

Abstract : Video de-noising is an important feature of many modern video encoding architectures since it can considerably enhance coding efficiency while at the same time improving objective and subjective quality. In this paper, a new scheme for video de-noising based on spatio-temporal filtering is presented, which can be used as a pre-processing stage within a video encoder. Spatial filtering is performing through the adaptive selection and combination of a wavelet based and a 2-D Wiener filter, while for temporal filtering we employ bi-directional block based motion compensation using the Enhanced Predictive Zonal Search (EPZS) algorithm. Experimental results are presented which show a significant improvement in video quality when the de-noised video sequence is encoded with an H.264 encoder.

Menu