Variable aperture panoramic imaging

Jing Lv, Qiang Zhao, Feng Dai, Yike Ma, Yongdong Zhang

Abstract: Panoramic imaging has important applications in virtual reality. Current panoramic imaging only can create a 360∘ panorama but cannot change the focal surface or support depth cues. In this paper, we present a variable aperture panoramic imaging method based on a two-sphere panorama model (TSPM). Based on this model, we propose a definition of Aperture for TSPM, which can precisely control the depth of field (DOF) of panoramas. With same aperture size and focus distance, two similar panoramas could be synthesized from a range of capture camera configurations. Extensive experiments have been carried out to validate our panoramic imaging method.