Next generation of shared virtual environment using tele-immersion

通信技术正在不断改变人们的互动方式，无论是商业还是个人。随着视频会议、网络会议和高速网络等技术的出现，通信从电子邮件和电话向前迈出了一大步。现在可以通过实时视频和音频进行通信，为电话的非实体语音带来了新的维度。随着网络技术的不断改进以及处理能力和显示能力的提高，新一代通信工具即将出现。共享虚拟环境为传统的视频会议提供了可能的继承者。共享的虚拟环境试图创建一个沉浸式的、共享的3D世界，该世界将人们跨网络连接起来，并且使他们不仅能够彼此交互，而且能够与各种其他形式的数字数据（例如，3D模型、视频、图像、文本和声音）交互。这个项目的目标是创造一个环境，使一个更现实，灵活和直观的方式进行沟通。为了促进自然的沟通，场景，特别是参与者必须尽可能真实地表示。新的低成本相机提供了以照片般逼真的实时方式查看真实的世界的机会，使其成为在共享VE中表示场景和化身的理想选择。 从多个摄像机网络中实时生成或合成场景的新视图是视图合成的问题，也是本研究工作的主要焦点之一。在这个研究项目中，我们将开发新的实时算法来执行视图插值以及前景/背景分割，以将参与者与他们的环境分开。我们还将开发新的服务器/客户端技术，能够处理现实生活中的网络问题，如延迟和智能管理，

Communications technology is constantly changing the way in which people interact, both for business and personally. With the advent of such technologies as video conferencing, web conferencing and high-speed networks, communicating has taken a huge step forward from email and telephone. It is now possible to communicate via real-time video and audio, bringing a new dimension to the disembodied voice of the telephone. With ever growing improvements to networking technology and increases in processing power and display capability, a new generation of communication tools are just around the corner. Shared virtual environments provide a possible successor for the traditional videoconference. Shared virtual environments attempt to create an immersive, shared, 3D world that connects people across networks and enables them to interact not only with each other, but with various other forms of digital data e.g. 3D models, video, images, text and sounds. The goal of this project is to produce an environment that enables a more realistic, flexible and intuitive way to communicate. To facilitate natural communication, scenes and particularly participants must be represented as realistically as possible. New low cost cameras provide the opportunity to view the real world in a photo-realistic real-time manner, making it an ideal choice for representing scenes and avatars in a shared VE. Creating or synthesizing in real-time these new views of a scene from a network of cameras is the problem of view synthesis and is one of the main focuses of this research work. In this research project, we will develop novel real-time algorithms to perform view interpolation as well as foreground/background segmentation to separate the participants from their environment. We will also develop new server/client technology capable of dealing with real-life networking problems such as latency and the smart management of