Data-dependent weighted average filtering for image sequence restoration

In this paper, the authors propose a data-dependent weighted average filter (Video-DDWA: Video Data-Dependent Weighted Average) aimed at restoration o f dynamic images deteriorated due to Gaussian additive noises. As proposed by the authors, this filter is based on the data-dependent processing, in w hich the filter weight is varied by multiple local information items derive d from the data proximal to the processing point in a still image, with ext ension of this processing from a spatial filter to a temporal-spatial filte r; and then the weight of adjacent frames is determined by detecting presen ce or absence of motion from the motion information such as new local infor mation. There are several conventional methods of restoration of dynamic im ages that involve motion compensation with subsequent spatiotemporal filter processing, but they all have limitations as to the filter restoration per formance due to noise-affected deterioration of the estimation accuracy of the motion vector. The proposed data-dependent filter using the motion info rmation has among others the following advantages: (1) Because it involves detection of the motion degree at which noise influence is suppressed, it e nables restoration of dynamic images featuring a high noise cancellation pe rformance in still areas, without motion deterioration due to filter proces sing in motion areas-in other words, with processing that does not cause de terioration of movement and (2) the computing load is lower than that with motion-compensated filters. As compared to the conventional methods using m otion compensation, this method enables attaining a high restoration perfor mance not only for signals with a low S/N ratio at which the accuracy of th e motion compensation estimation vector starts decreasing, but also for sig nals with a wide range of S/N ratios. The authors demonstrate with various application examples that this method is efficient for restoration of dynam ic images. (C) 2000 Scripta Technica.