Discrete bandlimited signal extrapolation by data segmentation

The authors present a technique that improves the computational efficiency of an algorithm used for discrete bandlimited signal extrapolation. It consists of extrapolating segments or subsequences of the original data set and then uncorrelating the individual results to yield a final estimate of the signal. It is found that the method is provides high-quality reconstructions while reducing the amount of computation required. However, it results in reduced resolution in the final reconstruction. Therefore, a tradeoff exists between computational speed and resolution in the final estimate