Xiaohong Jiang

MSPR: a retained-mode-based multiscreen parallel rendering system

Recent interest in large displays has led to the development of a variety of multi-screen display systems, most of which are comprised of multiple projectors situated for front or rear projection onto a screen to form a large logical display, and each projector is driven by one PC, augmented with high-performance graphics accelerator card. However, most of these systems are immediate-mode based, that is, all geometry data and control information must be sent on the run, and thus the interconnecting network bandwidth between clients and servers has become a fatal bottleneck. In this paper, we describe MSPR, a retained-mode based multi-screen parallel rendering system which offers the programmers with a OpenGL-like API. And the system could be run locally on one PF or parallelly rendered among any number of Computers. The details of data distribution, load balance, interconnecting network architecture, tile mosaic, etc. are hidden from the programmers.

A Parallel Framework for Interactive Rendering of Massive Complex Scenes on PCs Cluster

This paper present a parallel framework for interactively rendering massive complex scenes on PCs cluster. The framework integrates parallel rendering, scene graph, software acceleration techniques, and out-of-core data management. We propose a novel parallel rendering architecture based on dynamic rendering team and hybrid rendering pipeline to balance work load and to handle different graphics primitive distribution. A unified scene graph representation is designed for flexible object manipulation in upper applications and efficient rendering in lower rendering pipeline. We present a priority-based out-of-core data management scheme which guarantees fixed memory footprint and data scheduling efficiency. We have implemented this framework and developed two applications to test its performance on a cluster of 32 PCs: a walkthrough application for massive complex scenes and a distributed multi- player video game.

CoarseZ Buffer Bandwidth Model in 3D Rendering Pipeline

Depth traffic occupies a major portion of 3D graphics memory bandwidth. In order to reduce depth reading, we propose employing a low-resolution depth buffer, namely CoarseZ buffer, for tile-level depth culling before per-pixel test. The maximum depth of a tile is stored in the corresponding entry of CoarseZ buffer. Simulation results show that a small CoarseZ buffer can achieve remarkably high culling rate and significantly reduce z-reading bandwidth. We build a model that quantifies the influence of the CoarseZ design parameters on its efficiency and bandwidth. Test results of industrial benchmarks show that CoarseZ with tile size of 4 and bit depth of 16 can be a best selection to reduce memory bandwidth

Building high performance DVR via HLA, scene graph and parallel rendering

Distributed simulation and parallel rendering based on PC cluster have seen great success in recent years. To improve the overall performance, there is a trend to integrate modeling, simulation and visualization into a common distributed environment. In this paper, we propose a unified framework of building high performance distributed virtual reality (DVR) applications. The core components of this framework include the High Level Architecture (HLA), scene graphs and parallel rendering. The HLA supports interactive distributed simulation. Scene graphs are efficient to organize and manipulate scene data. And parallel rendering provides powerful rendering ability. This paper presents the in-depth architectural analysis of each components and derives a design that integrates them into a unified framework. Two DVR applications, including a remote navigation of massive virtual scenes and a multi-player video game, have been developed to evaluate the framework performance.

Predictive occlusion culling for interactive rendering of large complex virtual scene

We present an efficient occlusion culling algorithm for interactive rendering of large complex virtual scene with high depth complexity. Our method exploits both spatial and temporal coherence of visibility. A space hierarchy of scene is constructed and its nodes are rendered in an approximate front-to-back order. Nodes in view frustum are inserted into one of layered node lists, called layered buffers(LBs), according to its distance to the view point. Each buffer in the LBs is rendered with hardware occlusion queries. Using a visibility predictor(VP) for each node and interleaving occlusion queries with rendering, we reduce the occlusion queries count and graphics pipeline stalls greatly. This occlusion culling algorithm can work in a conservative way for high image quality rendering or in an approximate way for time critical rendering. Experimental results of different types of virtual scene are provided to demonstrate its efficiency and generality.

HIVE: a highly scalable framework for DVE

With the increasing requirements for distributed virtual environment (DVE): supporting larger number of participants and providing more smooth roaming and interactions, scalability is becoming a key issue. In this paper, we explore the scalability of participants and the scalability of servers, and mainly focus on three aspects: system architecture, communication model and interest mechanism. We present our middleware platform, HIVE, providing a variety of services such as data distribution, communication, event notification, etc. To achieve the reusability and interoperability of DVE applications, the interface specification of high level architecture (HLA) is employed as the reference. HIVE also contains the back-ends, which the middleware services depend upon. On HIVE, users can develop scalable DVE applications easily and quickly, concentrating on not the detail of distribution but the application logic. Finally an experimental demo on HIVE is given.

Parallel simplification of large meshes on PC clusters

Large meshes are becoming commonplace with the advance of 3D scanning, scientific simulation and CAD technology. While there are many algorithms proposed to simplify these large meshes, the time of simplification process is usually very long, especially for those algorithms based on iterative edge collapse. To address this problem, we propose two parallel schemes to speed up simplifying large meshes on a PC cluster. The first parallel simplification scheme partitions a large mesh into small sub-meshes, simplifies these sub-meshes in parallel in an in-core way and finally stitches the simplified versions together. The second scheme generates multiple mesh streams, applies stream simplification to them in parallel in an out-of-core way, and composes the final simplified mesh streams. We have implemented these two parallel simplification schemes and the experimental results show that our methods are able to speed up the iterative simplification of large meshes by a factor of 8 to 19 on a cluster of 24 PCs.

Research on Unified Object Model Supporting HLA-based Simulation and Parallel Rendering

There are two problems in the existing rendering platforms that support HLA-based simulation applications. One is the fact that the management of simulation entity objects and the management of rendering scene objects are separate, which leads to inefficiency in simulation and rendering. Another one is absence of rendering interface, especially for parallel rendering for users. In order to solve the above problems, we present a unified object model supporting HLA-based simulation and parallel rendering, which consists of heterogenous scene graph tree, action list and universal access interface. The Unified object model establishes an efficient data exchange bridge between HLA simulation and parallel rendering. Therefore massive complex scene and large quantity of simulation entities can be organized and managed in a unified form. An experimental demo is given at the end of this paper.

An experimental facial synthesis system using graph cut and gradient domain fusion

Based on the newly appeared image editing and image pro- cessing techniques, a novel interactive, computer-assisted system is proposed for facial synthesis. This paper presents the architecture of this facial synthesis system and gives a detailed description of the four key modules. The techniques used in these modules are also particularized. First, graph cut algorithm is used to automatically select region. Then gradient domain fusion is used to get better result. Finally, k-means method is used to improve efficiency. The experimental results show that our experimental facial synthesis system can produce visually good synthesized face images.

A scalable HLA-based distributed simulation framework for VR application

HLA-based Distributed Simulation technology is employed widely in Virtual Reality (VR) applications, such as military simulation, internet games, roaming etc. To support larger number of participants, scalability is becoming a key issue of VR applications. In this paper, we explore the characteristics of distributed simulation, and analyze the scalability of servers and participants, and classify our approach to improve scalability of VR applications into three aspects: a three-tier node management mode to accommodate more participants, an efficient management of servers to manage scalable number of clients and software development interface to achieve reusability and interoperability of VR applications. We present our middleware platform, HIVE, providing a scalable HLA-based distributed simulation framework for VR applications, on which users can develop VR applications easily and quickly. Then we give the method and view of application integration with HIVE. Finally an experimental demo is given.