Objective
Television channels, archivists and film producers all hold numerous archive programmes which, taken as a whole, make up a unique record of history, of artistic and cultural development and of all aspects of life in this century.
Each type of video and film format has its own share of particular degradation, with problems increasing with each replay. Several videotape recorders are no longer manufactured or serviced by the manufacturers ; this situation resulting in increased maintenance and operating costs.
As a result, archive holders must copy their collections onto new media, but the quality of the copy often remains well below today's standards, something which the general public, broadcasters and archive-based programme producers will be more and more reluctant to accept.
The economic exploitation of this type of programme depends a great deal on their quality level.
Solutions for restoring at an acceptable cost are therefore indispensable for the large-scale utilisation of television archives, thus contributing, through revenue generated, to the preservation of our audio-visual heritage.
The objective of AURORA is therefore to develop a fast and effective video restoration system with the following characteristics :
Real-time detection of impairments and estimation of quality level.
Restoration in real time with control of level of correction by the user.
Interactive restoration tools for high quality restoration or for badly damaged materials.
An in-depth understanding of the impairments linked to the video recording technology, and the methodologies and algorithms to address them.
A completely integrated set of techniques for exploiting the large stock of archive TV programmes : methodologies, algorithms, and real-time hardware.
The demonstrators will prefigure future products for the restoration of TV archives, and more generally, of impaired moving images programmes.
Expected Impact
The final aim of the project is to place at the disposal of European television channels, programme producers and educators, a service that will enable them to make much more effective use of video and film assets (several million hours). At the same time, this should free up resources for conserving, broadcasting and restoring rare and irreplaceable documents.
Example of source and restored sequence
Main contributions to the programme objectives:
Main deliverables
An enhanced restoration system for film archives.
Contribution to the programme
To build and demonstrate an integrated prototype system to cost-effectively restore material from film and video archives in real-time.
Technical Approach
Preliminary work consists of making a detailed evaluation of various videographic documents, identifying and understanding defects, testing and developing new and existing restoration algorithms and in selecting functions to be implemented in real time
The system will process in real time :
video noise, film grain and other random continuous defects,
impulsive and erratic impairments (video drop-outs, film dirt, sparkle etc.),
geometric instability,
colorimetric instability,
continuous linear defects (streaking, echoes, loss of detail etc.).
However, the final decision may always be taken by the operator who may, at any moment, decide to change a parameter, an action, a particular time-code, a particular position, to manually point out a zone to be corrected, or even at times to modify the editing of the programme.
Advanced software tools will be developed for this purpose, enabling the trained operator to obtain a good quality image in a minimum amount of time by intensive use of the real-time processing resources.
Specific research work, based on similar algorithms, is underway, into the extraction of photographic quality images from video programmes. This work will provide a complete software solution. The system will be simple enough to be implemented by non-specialist users, for example journalists.
Summary of Trial
The universal exhibition of 1998 in Lisbon (Expo'98), will provide an ideal opportunity for demonstrating the performance of the system.
Several hours of television archive documents will be restored with help from the new system. Documents dealing with the ocean, the theme of the exhibition, will be selected. Thanks to the new system, these will be subjected to in-depth processing and will be put on view during the exhibition together with on-site demonstrations of restoration. In these presentations, a trained operator will be using the system on actual archives, thus proving the effectiveness of the system.
Key Issues
Video processing speed and efficiency
Real-time robust motion estimation techniques with respect to additive and replacement noise
Real-time motion-compensated noise reduction
Real-time motion-compensated impulsive noise concealment
Artefacts understanding, detection, and logging
User interfaces and hardware for high-level restoration
Still pictures extraction from video
Fields of science
Topic(s)
Call for proposal
Data not availableFunding Scheme
CSC - Cost-sharing contractsCoordinator
94366 Bry sur Marne
France