Découverte automatique de la région d'intérêt en recalage d'images direct

The goal of the direct aggregative image registration problem is to aggregate in an incremental way a set of images by minimizing the pixel intensity discrepancy between the current aggregate and each new image. One of the bigest drawback of this method is that it relies on an a priori knowledge of the overlapping area (or region of interest) between the aggregate and the new image. In practice, the region of interest is specified by the user. In this article, we propose and compare several new approaches for which an a priori knowledge of the region of interest is not required anymore.     «

The goal of the direct aggregative image registration problem is to aggregate in an incremental way a set of images by minimizing the pixel intensity discrepancy between the current aggregate and each new image. One of the bigest drawback of this method is that it relies on an a priori knowledge of the overlapping area (or region of interest) between the aggregate and the new image. In practice, the region of interest is specified by the user. In this article, we propose and compare several new...     »