Multihypothesis recursive video denoising based on separation of motion state

A multihypothesis recursive video denoising filter (MRF) based on separation of motion state is proposed. For video sequence degraded by additive Gaussian white noise, local motion state will be detected combining multiple hypotheses (temporal predictions) first. Then different denoising schemes will be selected to suppress the noise according to the local motion state. Areas detected as stationary motion will be filtered by multihypothesis motion compensated filter (MHMCF), whereas areas detected as non-stationary motion will be filtered by self-cross-bilateral filter (SCBF). The definitions of stationary motion state and non-stationary motion state are given. In addition, the threshold used to classify motion state is equal to the noise standard deviation. The simulation results show that MRF outperforms conventional denoising methods like joint filtering scheme, spatio-temporal varying filter and MHMCF both in peak signal-to-noise ratio and visual quality.