Superresolution through error function extrapolation

Summary form only given, as follows. The problem of achieving superresolution beyond the diffraction limit has been approached using the constrained iterative algorithm to reduce the error energy of the limited function. However, it is also possible to use the algorithm with complementary constraints to find the total error function. Subtracting this from the limited function yields the superresolved function. One cycle of the iterating process is representable as a matrix which, when multiplied by the correct error function, gives back the same function. It proves to be very easy and instructive to solve this eigenvector problem. In so doing it is possible to control the upper limit of the bandwidth of the found error function and to trade bandwidth for noise performance. In an N-element image with A sample points in the support region, this approach appears to be equal to using the impulse response function N times to sharpen the image. That is, it uses all of the data available in the bandlimited image. However, it is not clear whether this method gives the maximum noise-bandwidth recovery performance