Role of wavelet transforms in image restoration

Image processing techniques including image restoration rely heavily on discrete Fourier transform (DFT) for frequency domain representation for analysis of its frequency content. Even though wavelet analysis has been around for more than a decade, much of its potential as a tool to analyze time-frequency localization of signal has not been properly tapped. In blind iterative deconvolution where a degraded image is restored with minimum a priori information about the original image or the point spread function (PSF), it is almost impossible to evaluate whether a restoration is achieved or not without human observation. We propose a new approach using wavelet decomposition to assess an image being restored or not automatically.