In this article, we propose a depth map refinement method that increases the quality of immersive video. The proposal highly enhances inter-view consistency maps (estimated or acquired by any method), crucial for achieving required fidelity virtual view synthesis process. described method, only information from is used, as use texture can introduce errors in refinement, mostly due to color inco...

Novel-View Human Action Synthesis aims to synthesize the movement of a body from virtual viewpoint, given video real viewpoint. We present novel 3D reasoning target first estimate mesh and transfer rough textures 2D images mesh. As this may generate sparse on due frame resolution or occlusions. produce semi-dense textured by propagating transferred both locally, within local geodesic neighborho...

This paper addresses the problem of novel view synthesis by means neural rendering, where we are interested in predicting at an arbitrary camera pose based on a given set input images from other viewpoints. Using known query and poses, create ordered observations that leads to target view. Thus, single is reformulated as sequential prediction task. In this paper, proposed Transformer-based Gene...

In this paper, we tackle the problem of ground-view panorama image generation conditioning on top-view aerial image, which is a challenging due to large gap between domains associated with different view-points. Instead learning underlying cross-view mapping by feedforward network as previous methods, propose novel adversarial feedback GAN framework named PanoGAN consisting two key components: ...

One component of the Virtual Square framework[12, 3, 4], is View-OS. Traditional OS’s implement the global view assumption: any two processes running on a computer share the same view of their execution environment. This “global” view includes the meaning of pathnames, network stacks, routing rules, devices etc. While processes may have different permissions to access these resources, the namin...

In this paper a new "range-space" approach, for rendering visual models using a network of multiple omnidirectional vision sensors (ODVS) is presented. This integrated approach allows for simultaneous extraction of 3-D range as well as visual models. The approach requires three distinct steps of analyzing multiple ODVS video input streams: 1) Search, 2) Match, and 3) Render. At the output, a us...

This paper presents an implemented system for automatically producing prosodically appropriate speech and corresponding facial expressions for animated, threedimensional agents that respond to simple database queries in a 3D virtual environment. Unlike previous text-to-facial animation approaches, the system described here produces synthesized speech and facial animations entirely from scratch,...