Linear filtering of spatially invariant image sequences for feature separation under three types of image noise

Many important imaging applications in medical imaging, remote sensing and other areas result in a set of images of the same scene with no relative scene-sensor motion and in which the pixel intensities are the linear sum of the contributions of the distinct features in the scene. Such image sets are called linearly additive, spatially invariant (LA SI) image sequences. Previous research has shown, both mathematically and with examples, that a K-image sequence with M distinct features (M\n\n\t\t