Fast iterative methods for high-resolution multispectral image restorations
Dr Ng, Kwok Po (Principal investigator)
Professor Chan Raymond Hon Fu (Co-Investigator)
RGC General Research Fund (GRF)
HKU Project Code
General Research Fund (GRF)
High-resolution color/multispectral image restoration problems arise in a variety of scientific, medical and engineering applications: from the daily-life TV broadcasting to the sophisticated satellite remote sensing. The problem is to reconstruct high-resolution images from low-resolution frames. The resulting linear systems are usually Toeplitz-like systems. This research is centered on the development of fast iterative methods for such systems. For single-spectral (i.e. gray-level) images, we have solved the systems by the preconditioned conjugate gradient method with cosine transform-based preconditioners and the method is faster than any existing ones. The purpose of this research is to develop and implement this fast iterative method for color/multispectral image restorations; and to establish theoretically the convergence rate of the method.