engleski

Algorithm for Combined Wavelet Quasi-Superresolution

This article gives an outline of the novel algorithm for image superresolution based on wavelets. In this algorithm morphological operations in low-resolution image are combined with approximation of high-resolution wavelet coefficients. Since it involves only a single original low-resolution image, we call it quasi-superresolution. The final result is high-resolution image obtained from a single source image applicable in analysis not only of standard superresolution problems, but in stationary scenes and biomedical imaging as well. In one step, four pixels instead of a single LR pixel are obtained by the algorithm. Algorithm can be used iteratively.

wavelet; restoration; approximation; quasi-superresolution; morphological operations; image processing

nije evidentirano