A robust image registration method based on total variation regularization under complex illumination changes

Image registration is the process of aligning two or more images of the same scene which may be taken at different times, from different viewpoints or by different sensors. Lots of novel approaches have been developed for this purpose. They can be classified into two categories: feature-based and area-based methods [1 –4]. The feature-based approaches consist of 3 steps: (1) extracting features from each image, (2) finding the correspondences between the extracted features using a similarity metric, and (3) estimating the transform parameters according to the locations of the corresponding points.
Source: Computer Methods and Programs in Biomedicine - Category: Bioinformatics Authors: Source Type: research