Image Registration for the Alignment of Digitized Historical Documents
This work addresses the alignment of historical documents for archivists and researchers, but it is incremental as it applies existing registration methods to a specific domain.
The paper tackled the problem of aligning digitized historical documents using hyperspectral image registration, evaluating various algorithms and selecting an intensity-based method with cubic B-spline transformation and similarity measures like residual complexity and localized mutual information, which showed acceptable performance in handling challenges such as non-rigid deformations and intensity distortions.
In this work, we conducted a survey on different registration algorithms and investigated their suitability for hyperspectral historical image registration applications. After the evaluation of different algorithms, we choose an intensity based registration algorithm with a curved transformation model. For the transformation model, we select cubic B-splines since they should be capable to cope with all non-rigid deformations in our hyperspectral images. From a number of similarity measures, we found that residual complexity and localized mutual information are well suited for the task at hand. In our evaluation, both measures show an acceptable performance in handling all difficulties, e.g., capture range, non-stationary and spatially varying intensity distortions or multi-modality that occur in our application.