Adaptive Error Concealment With Radial Basis Neuro-Fuzzy Networks For Video Communication over Lossy Channels

摘要

Error is inevitable in data trans-receiving over communication channels and its concealment is indispensable for standard codecs to enhance their quality of services. In recently standard video codecs, error resilience is based on absolute spatial and temporal approaches being applied uniformly for several types of video sequences. In this paper, an adaptive error concealment method based on radial basis neuro-fuzzy networks is proposed. Simulation results show that the proposed algorithm can gain up to 5 dB compared to the standard median algorithm