Desert island dreaming
18 November 2003
Imagine being anywhere in the world and instantaneously being able to call up a piece of film and watch it on a portable device that provided images of unparalleled quality. Turning this into reality is the dream behind 3C Research, a University-based company.
The city of Bristol has long been renowned for its activities in TV production and high-tech electronics. In 3C Research, world-class expertise developed at the University will join forces with a range of industrial partners, including Granada, Toshiba and Qinetiq, to look into the technological future. Recently awarded £7.6m by the Department of Trade and Industry to progress its adventurous research programme, 3C Research will bring together a dynamic mix of academic research excellence, production creativity, and technological know-how. With the industrial partner support, funding available for the programme is over £11m in the first three years.
Increasingly people are looking to be liberated from sitting at a desk in order to access information, but this requires significant advances in equipment portability and access to interactive information, as well as the applications that support them. The dramatic advances in multimedia services and the new, user-friendly devices needed to deliver the dream will build upon five projects that arose out of research originally done in the University’s Electrical and Electronic Engineering and Computer Science departments.
Generating realistic animation is an expensive and highly skilled process. Motion Ripper aims to provide animators in the broadcast and games industries with a suite of tools that will enable them to produce complex, realistic animation quickly and easily.
Access to Granada’s online video database will provide the project with an extensive range of motions to sample. Ultimately, using the software being developed, animators will be able to extract (rip) motion data from the reference video which can then be used in a computer graphics scene to drive character motion. Unique versions of these motion ‘signatures’ will then be automatically generated, allowing the animators to extend clips or create complex effects. For example, the motion signature of a bird flying can be translated to create a plane that flaps its wings.
As well as developing the software necessary to enable these effects, an online database of motion signatures will be produced by Matrix Data who bring a high level ‘semantic’ approach to metadata cataloguing and classification. Metadata is the descriptive information about the data and the semantic approach uses a global classification system that can be processed by anyone in any way. The final outcome of the Motion Ripper project will be online databases that can be rapidly searched for audio clips and images to enhance the entertainment industry.
ICBR – Intelligent Content-Based Retrieval
If a producer does not have the funding or time to go to South Africa and film a lion chasing zebras, how can he find such a clip amongst all the film footage available? Television and film producers have a real need to be able to organise, store and retrieve vast amounts of media images. Furthermore, their requirements are growing as more and more media becomes available. The ICBR project proposes to solve the problem by combining a large-scale multimedia server and database with intelligent shot-recognition and retrieval software.
When completed, the ICBR demonstrator will include a set of software tools for the automatic recognition of images and audio, and for the shot selection and editing of television programmes. The system will select video and audio material according to pre-set criteria, which will include image-similarity measures and object recognition. The selected media will then be presented in an ordered way ready for incorporation into an editing system in order to construct new sequences or complete programmes.
RoD – Rendering on Demand
The computer graphics industry, and in particular those involved with films, games and virtual reality, continues to demand more and more realistic computer-generated images. Despite the ready availability of modern, high-performance graphics, the complexity of the content being modelled and the high fidelity required of the images means that rendering such images (a technique for producing images from their three-dimensional geometry) is simply not possible in a realistic timeframe on a single computer.
RoD aims to provide a high-performance, high-fidelity rendering system that will be accessible from many locations. This will enable users to submit their content to the system and see the desired images rendered in a reasonable, and ultimately realtime, time-frame. RoD will achieve this goal of ‘realism in real-time’ by combining parallel processing – a number of computers working in tandem – with visual perception techniques. These utilise the fact that the human eye fails to notice certain aspects of an image. By exploiting our understanding of the human visual system, significant time can be saved by simply not rendering those parts we will not notice.
The motion signature of a bird flying can be translated to create a plane that flaps its wings
OSIRIS – Open Infotainment Services In Radio Interconnected Systems
With the Government aiming to extend the information age to everyone, a key requirement will be fast, flexible and reliable broadband access if the opportunities and benefits of a knowledge-based economy are to become inclusive, and the promise of ‘access to the internet for everyone who wants it by 2005’ is to be realised. Wireless technology is widely regarded as a means of enabling this revolution, however significant research and development is still necessary before wireless can offer ‘wired quality’ at a realistic price for the mass market.
OSIRIS aims to address this need with the development of novel technologies that will provide an adaptable and resilient radio infrastructure. This will give everyone, everywhere, at any time, seamless access to top-quality broadband internet facilities via handheld devices that tap into a network of public hotspots.
ROAM4G – Robust and Scaleable Multimedia Content Delivery over Optimsed 4G Communication Networks
Future wireless networks will also support high-speed communications in dense urban and other environments. However, the delivery of video to a wireless device in a mobile environment remains a challenging problem. ROAM4G will provide a robust and scaleable coding solution for high quality multimedia streaming and conversational/interactive services. Application areas for this include video enhanced services for mobile devices such as mobile phones, video conferencing, home networks for broadcast distribution, image and video database retrieval systems such as OSIRIS.
Since the human eye fails to detect certain aspects of an image, significant time can be saved by not reproducing those parts we will not notice
The three Cs in 3C Research stand for communications, computing and content. When combined in the exciting ways envisaged by 3C Research, these words will impact on the economy and people’s lives in ways hitherto unimagined – even on a desert island.