Abstract
Superresolution theory and techniques seek to recover signals from samples inthe presence of blur and noise. Discrete image registration can be an approachto fuse information from different sets of samples of the same signal.Quantization errors in the spatial domain are inherent to digital images. Weconsider superresolution and discrete image registration for one-dimensionalspatially-limited piecewise constant functions which are subject to blur whichis Gaussian or a mixture of Gaussians as well as to round-off errors. Wedescribe a signal-dependent measurement matrix which captures both types ofeffects. For this setting we show that the difficulties in determining thediscontinuity points from two sets of samples even in the absence of othertypes of noise. If the samples are also subject to statistical noise, then itis necessary to align and segment the data sequences to make the most effectiveinferences about the amplitudes and discontinuity points. Under some conditionson the blur, the noise, and the distance between discontinuity points, we provethat we can correctly align and determine the first samples following eachdiscontinuity point in two data sequences with an approach based on dynamicprogramming.