A comparative study of state-of-the-art evolutionary image registration methods for 3D modeling

Image registration (IR) aims to find a transformation between two or more images acquired under different conditions. This problem has been established as a very active research field in computer vision during the last few decades. IR has been applied to a high number of real-world problems ranging from remote sensing to medical imaging, artificial vision, and computer-aided design. Recently, there is an increasing interest on the application of the evolutionary computation paradigm to this field in order to solve the ever recurrent drawbacks of traditional image registration methods as the iterated closest point algorithm. Specially, evolutionary image registration methods have demonstrated their ability as robust approaches to the problem. Unlike classical IR methods, they show the advantage of not requiring a good initial estimation of the image alignment to proceed. In this contribution, we aim to review the state-of-the-art image registration methods that lay their foundations on evolutionary computation. Moreover, we aim to analyze the performance of some of the latter approaches when tackle a challenging real-world application in forensic anthropology, the 3D modeling of forensic objects.

► Image registration aims to find a transformation between two different images.
► Evolutionary computation uses computational models of evolutionary processes.
► Evolutionary computation has been widely applied in image registration.
► We review the state-of-the-art methods that are based on evolutionary computation.
► We study the evolutionary algorithms performance in 3D modeling of forensic objects.