Name
Title | ||
|---|---|---|
Robust image registration with illumination, blur and noise variations for super-resolution |
Abstract | ||
|---|---|---|
Super-resolution reconstruction algorithms assume the availability of exact registration and blur parameters. Inaccurate estimation of these parameters adversely affects the quality of the reconstructed image. However, traditional approaches for image registration are either sensitive to image degradations such as variations in blur, il- lumination and noise, or are limited in the class of image transfor- mations that can be estimated. We propose an accurate registration algorithm that uses the local phase information, which is robust to the above degradations. We derive the theoretical error rate of the estimates in presence of non-ideal band-pass behavior of the filter and show that the error converges to zero over iterations. We also show the invariance of local phase to a class of blur kernels. Ex- perimental results on images taken under varying conditions clearly demonstrates the robustness of our approach. |
Year | DOI | Venue |
|---|---|---|
2008 | 10.1109/ICASSP.2008.4517856 | ICASSP |
Keywords | Field | DocType |
image reconstruction,image registration,blur kernel,blur parameter,blur variation,illumination variation,image degradation,image transformation,noise variation,parameter estimation,robust image registration,super-resolution reconstruction algorithm,Local Phase,Registration,Super-Resolution | Iterative reconstruction,Computer vision,Invariant (physics),Computer science,Word error rate,Gaussian blur,Robustness (computer science),Artificial intelligence,Estimation theory,Image restoration,Image registration | Conference |
ISSN | Citations | PageRank |
1520-6149 | 0 | 0.34 |
References | Authors | |
9 | 3 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Himanshu Arora | 1 | 77 | 9.54 |
| Anoop M. Namboodiri | 2 | 255 | 26.36 |
| C. V. Jawahar | 3 | 1700 | 148.58 |