Nonlinear model-based spatio-temporal filtering of image sequences

A Volterra model-based spatio-temporal filter for the enhancement of noise-corrupted image sequences is considered. This model uses estimates of higher-order statistics (HOS) to filter non-wide-sense stationary (WSS) image sequences that cannot be correctly modeled by second-order statistics alone. Some results are shown for this filter when it is applied along the direction of motion in image sequences with simulated noise.<
>