Eigenspace of Spinning Panoramic Images Ales Leonardis Computer Vision Laboratory Faculty of Computer and Information Science University of Ljubljana, Slovenia Recent studies have shown good prospects for appearance-based localization of mobile robots using panoramic sensors. Panoramic sensors capture a wide field-of-view and enable efficient characterization of a location by a single panoramic image. The redundant information captured in similar views is efficiently handled by the eigenspace approach. In this talk, I will present some interesting properties of eigenspaces of inplane rotated panoramic images, which enable an efficient eigenspace construction.