Research


Multimedia Networking and Video Streaming
Interactive Multiview Video
Video Analysis and Video Mining
Embeded Medeia Processing and SoC
Video Coding for Heterogeneous Network and Wireless Network
Digital Rights Management

 

Multimedia Networking and Video Streaming

Back to top

ADVANCES in computers, networking, and communications have created new distribution channels and business opportunities for the dissemination of multimedia content. Streaming audio and video over networks such as the Internet, local area wireless networks, home networks, and commercial cellular phone systems has become a reality and it is likely that streaming media will become a mainstream means of communication. Despite some initial commercial success, streaming media still faces challenging technical issues, including quality of service (QoS) and cost effectiveness. For example, the emerging streaming architectures such as peer-to-peer (P2P) networks and wireless ad hoc networks have also raised many interesting research challenges.

The goal of the group is to research and develop advanced and efficient techniques for video streaming delivery. To achieve the goal, various networking and multimedia delivery issues will be addressed. We also develop a peer to peer based live video streaming system to prove and demonstrate the concepts of our results.

Multicast is a fundamental and important mechanism to deliver scalable multimedia contents. Unfortunately the Internet nowadays still cannot support global multicast. In recent years, peer-to-peer (p2p) based multimedia streaming has aroused much interest both in research communities and in industries. In a p2p streaming system, multimedia contents should be delivered to a large pool of distributed users with low delay and high quality. Fueled by advances in networking and compression technologies, p2p multimedia streaming has experienced initial deployment success for applications such as Internet TV, video conferencing, and surveillance. A scalable peer-to-peer multimedia system should support many hosts, possibly in excess of hundreds or even millions, with diverse heterogeneity in bandwidth, capability, storage, network, and mobility. It should also be able to support various applications and file formats under dynamic user arrival and departure, frequent host failures and unavailability, and unpredictable network traffic and congestion. To achieve these goals, it is particularly important to address the challenges in architecture design, network/transport support, resource discovery and content delivery mechanisms.

Therefore, the group has designed and implemented two peer-to-peer based live video streaming systems, providing service over campus network and global Internet. For the first experience, a multi-sender based hierarchical explicit overlay structure was deployed to demonstrate the scalability of peer to peer concept. Then, to address the problems of single point failure and dedicated server workload, the group advocated gossip-based distributed membership management and with the help of efficient ‘push+pull' streaming mechanism, the system has been deployed by CCTV for live broadcasting Spring Festival Evening Show. Most excitingly, it was evidenced this architecture could scale to reliably support as many as 15,239 concurrent users with support of only one common streaming server.

Now the group proceeds ongoing research work along several arenas, including nodes heterogeneity and scalable coding, media on demand service, efficient video coding schemes, secure multicast and incentive mechanisms

 
Interactive Multiview Video Back to top
The next generation multimedia applications will be interactive and realistic based on network. In one hand, people want to be more active and participant to meet the requirement of personality, in the other hand, they hope to get realistic experience while viewing the scene or interacting with other people. This project is proposed to resolve the represent, interaction, codec and transmission of 3D interactive video, and implement the interactive network multimedia application in the near future with the stereo video and multi-view video technology putting into practicability.

Objects:
   Viewpoint selecting and multi-view 3D image synthesizing
   High efficiency multi-view video codec algorithms
   Streaming technology for Multi-view video
   A prototype of multi-view video streaming system
 
Video Analysis and Video Mining Back to top

Digital videos are universal in nowadays. However, for the lack of efficient ways in management and retrieval, increasing video data are getting overwhelming. Efficient video indexing and retrieval have emerged as a critical and challenging problem in multimedia applications. Various features such as color, shape, texture and motion are being used for video indexing and retrieval. In the aforementioned approaches, however, the performance of video retrieval is far from satisfactory due to the gap between low-level features and high-level semantics. From the point of view of users, instead of querying by low-level similarity and examples, semantics based video retrieval is the nature and efficient manner. Therefore, recent work in video indexing and retrieval has witness an interesting shift from traditional low-level feature based approaches to high-level semantic analysis.

Video Mining can be defined as the unsupervised discovery of patterns in audio-visual content. Video patterns are repetitive segments in a video stream with consistent statistical characteristics. Such segments can often be interpreted in relation to distinctive semantics or concepts, particularly in structured domains like sports game and news report.

While much work in the literature explores the supervised learning techniques that are capable of learning the target structure once the domain knowledge is encoded in the training data, the choice of the feature set, the construction of the statistical model, and the design of the learning algorithms. Unsupervised mining algorithms, on the other hand, tries to find statistical descriptions of the structures with much less information available, aiming at alleviating the burden of labeling and training, as well as providing a scalable solution for generalizing video indexing techniques to heterogeneous content collections.

 
Embeded Medeia Processing and SoC Back to top

eMeeting是一个基于嵌入式平台的实时视频通讯原型系统,硬件平台基于 Intel XScale PXA255 处理器,软件基于 Linux 等各种开放源码软件,实现了全软件的符合 H.263 视频编解码标准的视频通讯原型系统,在同时运行 QCIF 大小的视频编码、解码、无线网络通讯、 LCD 显示等功能模块的条件下,系统的视频通讯速率达到了 8/fps 。后续我们将在 PXA27X 硬件平台和 H.264 编码标准上开展进一步的研究。

多媒体处理功能将是未来个人终端设备上的重要功能之一,但是多媒体处理计算量非常高,这对于以电池供电、计算资源有限的嵌入式终端设备来说是一个很大 的挑战。本研究以多媒体移动终端为典型研究对象,希望通过嵌入式设备上的软件优化技术、计算迁移技术、构件化软件技术和可重构硬件技术,在无需用户干预的 情况下,实现在嵌入式终端设备上流畅地运行未来以多媒体处理为代表的各种复杂应用。

 
Video Coding for Heterogeneous Network and Wireless Network Back to top
 
 
Digital Rights Management Back to top

The subject focuses on various security technologies and solutions used for Digital Rights Management including digital watermarking, rights expression language, trust management, license management etc.

With the explosive development of network community and media culture, digital content offers an increasing number of opportunities for people. However, the ability for anyone to make perfect copies and the ease by which those copies can be distributed also facilitates misuse, illegal copying and distribution, plagiarism, and misappropriation. There emerges great desire for Digital Rights Management (DRM) systems which can protect the economic value of digital contents and owners' rights.