Image registration using PCA and gradient method for super-resolution imaging
So Sasatani, Xian‐Hua Han, Yen‐Wei Chen · International Conference on Software Engineering · 2010
Super-resolution (SR) enhancement from multi-frame low-resolution (LR) images (multi-frame super-resolution) has been a well-studied topic in the literature. Image registration is the most important part for multi-frame super-resolution, and accurate alignment of LR images would contribute a critical role for the final success of SR image reconstruction. In this paper, we propose to combine the Principle Component Analysis (PCA) based registration method, which can perform object alignment in real-time and without constraints on the three registration parameters (i.e., translation, rotation, and scaling), and gradient registration method, which can perform precise registration with minor image movement. Experimental results show that the reconstruction SR images by our proposed method have much higher quality than those by the state of art algorithms.