New approaches for 3D streaming and rendering based on predictive modeling

Abstract

The Interactive 3D Computer Graphics has gained attention over the network in recent times through various 3D Networked Virtual Environments such as Massively Multi user Online Games Virtual Universities Virtual Museums and 3D Art Galleries These 3D networked virtual environments comprise of a set of 3D objects characterized by high resolution meshes On demand the client has to wait longer until these high resolution meshes are completely downloaded for every request In order to reduce this waiting time client server frameworks have been defined to progressively stream visible portion of the 3D content based on user interaction pattern predictions The streamed 3D content is newlinerendered on to a client machine Thus in addition to the idea of progressive streaming and rendering of visible meshes the anticipatory streaming of additional meshes in advance based on the user interaction pattern is proposed In this work anticipatory streaming model is referred as Predictive Streaming Predictive Streaming would thus help in effective utilization of the network bandwidth client memory and reduce rendering latency In this thesis the mesh attributes of 3D models which carry the dominant features for 3D reconstruction are considered for implementation and analysis of streaming and rendering The meshes which are triangular geometry shapes of newlinethe 3D objects comprise of vertices and its connectivity newline newline