This paper presents a method for view synthesis from multiple views and their depth maps for free navigation in Virtual Reality with six degrees of freedom (6DoF) and 360 video (3DoF+), including synthesizing views corresponding to stepping in or out of the scene. Such scenarios should support large baseline view synthesis, typically going beyond the view synthesis involved in light field displays [1]. Our method allows to input an unlimited number of reference views, instead of the usual left and right reference views. Increasing the number of reference views overcomes problems such as occlusions, tangential surfaces to the cameras axis and artifacts in low quality depth maps. We outperform MPEG's reference software, VSRS [2], with a gain of up to 2.5 dB in PSNR when using four reference views.Index Terms -View synthesis, depth image based rendering, free navigation

No abstract

No abstract

No abstract

A RGB-D image combines, for each pixel, the classical three color channels with a fourth channel providing depth information. Devices that produce RGB-D images in real time with a rather good resolution are currently available on the market. With this type of device, it is possible to acquire and to process, in real time, 3D textured information, paving the way for numerous applications in the field of computer imaging and vision.In this paper, we analyse the accuracy of a low cost system and we see how this kind of device and the RGB-D images it produces allow us to acquire 3D models of real objects. A first application is presented that combines multiple RGB-D images of a static scene, taken from different viewpoints, in order to reconstruct a complete 3D model of the scene. A second application combines on-the-fly RGB-D images coming from multiple devices, generating a 3D model where the problems of occlusions inherent in monocular observations are drastically reduced.