Generation of Omnidirectional Image Without Photographer.

In order to create a virtual reality (VR) space using omnidirectional images, it is desirable to use images without the photographer's inclusion. In this study, we propose a method to generate an omnidirectional image without the photographer's inclusion by using multiple images taken by an omnidirectional camera. In the proposed method, the photographer rotates around the omnidirectional camera and takes several images. Next, we perform feature point matching on the omnidirectional images and unify the appearance of all the images by using the amount of translation calculated from the matching. Finally, the images are combined with graph cut and Poisson image editing to produce an omnidirectional panoramic image without the photographer in it.