3D image and video technology

Abstract

3D imaging and video technologies are of growing interest in recent times because of having potential applications in many fields such as robotics, visualization, 3DTV, autonomous vehicles, driver assistance, “flying eyes”, intelligent human-machine interfaces, and so forth. To advance those technologies, computer vision and pattern recognition, which originated separately, are today increasingly interacting. In fact, papers published in this special issue address areas of both computer vision and pattern recognition. After going through a rigorous anonymous peer reviewing process and several revisions, seven papers are finally presented in this special issue. The first paper by Kobayashi, Sakaue and Sato, entitled “Multiple view geometry of projector-camera systems from virtual mutual projection” presents a calibration method for projector-camera systems. The proposed method generates virtual mutual projections between projectors and cameras by considering the shadow of cameras and the shadow of projectors and then discusses multiple view geometry for calibration. This work advances existing multiple view geometry so that it can also deal with projector-camera systems. The second paper by Jarusirisawad, Hosokawa and Saito, entitled “Diminished reality using plane-sweep algorithm with weakly-calibrated cameras” addresses an on-line method for generating free viewpoint views using captured images at different viewpoints. In this paper, the plane-sweep algorithm already proposed by the authors is extended from in the Euclidean space to in the projective space. The advantage of this method exists in excluding occluding objects for view synthesis. In “Object segmentation under varying illumination: stochastic background model considering spatial local-