Abstract
Super-resolution algorithms have demonstrated impressive image restoration results in simulations and some practical applications. Analysis of these algorithms' restoration performance is always a concerned problem of application. Many factors from the object, image acquirement process and restoration algorithm have their effects on the performance of super-resolution. The object of this paper is one-dimensional signal super-resolution restoration, which principally analyze some main effects on super-resolution: the spatial extent of the object, point-spread function, sampling rate, noise level and a priori constraints. It is shown that super-resolution performance is inversely related to the spatial extent of the object and the noise level of the image. For sampled data, restoration performance is related to the difference between sampling rate and the system cutoff frequency. These one-dimensional results can easily be extended to two dimensional images, and may help people to improve in existing super-resolution algorithms.
| Original language | English |
|---|---|
| Pages (from-to) | 816-820 |
| Number of pages | 5 |
| Journal | Beijing Ligong Daxue Xuebao/Transaction of Beijing Institute of Technology |
| Volume | 24 |
| Issue number | 9 |
| Publication status | Published - Sept 2004 |
Keywords
- Image restoration
- Maximum a posteriori (MAP)
- One-dimensional signal
- Super-resolution