High-resolution images are often desired but made impossible because of hardware limitations. For the high-resolution model proposed by Bose and Boo, the iterative wavelet-based algorithm has been shown to perform better than the traditional least square method when the resolution ratio M is two and four. In this paper, we discuss the minimally supported biorthogonal wavelet system that comes from the mathematical model by Bose and Boo and propose a wavelet-based algorithm for arbitrary resolution ratio M ≥ 2. The numerical results indicate that the algorithm based on our biorthogonal wavelet system performs better in high-resolution image reconstruction than the wavelet-based algorithm in the literature, as well as the common-used least square method.
ASJC Scopus subject areas
- Signal Processing
- Electrical and Electronic Engineering