UPF is leading a European project to improve energy efficiency and savings by the media and audiovisual sector through artificial intelligence

The Emerald project is driven by an interdisciplinary consortium of seven European partners, coordinated by UPF and comprising Trinity College Dublin, the BBC and the following four technology providers: Brainstorm Multimedia, Disguise Systems, FilmLight GmbH, and MOG Technologies. Given the huge increase in digital and audiovisual information and entertainment content, the project, funded by the EU's Horizon programme, aims to make these sectors more sustainable.

UPF will be coordinating the European project, Emerald, to develop artificial intelligence (AI) applications to make the production processes of digital media, the audiovisual sector and digital entertainment more efficient and reduce their level of energy expenditure. Faced with the strong increase in the production of audiovisual and augmented reality content in recent years, the Emerald project seeks to help make these sectors more sustainable.

The interdisciplinary consortium promoting the Emerald project, made up of six European partners besides UPF, is being coordinated by Josep Blat and Coloma Ballester. Blat and Ballester are the coordinators of the GTI (Interactive Technologies Group) and IMVA (Intelligent Multimodal Vision Analysis) research groups, respectively, both at the UPF Department of Information and Communication Technologies (DTIC).

Also members of the Emerald consortium are the BBC, Trinity College Dublin and four companies supplying the necessary technologies to develop media or audiovisual products (Brainstorm Multimedia, Disguise Systems, FilmLight GmbH, and MOG Technologies).

Earlier this March, the European Union's Horizon programme granted 3.1 million to develop the Emerald project, which will research and develop AI applications, based on Machine Learning technologies for the media and audiovisual sector. The two-and-a-half-year project will begin in October.

The coordinator of the GTI (DTIC-UPF), Josep Blat, summarizes the reasons that have led to the Emerald project: Film, television, series consume a lot of energy, both in the creation and in the consumption of audiovisual products, and artificial intelligence, based on processing a lot of data, also accounts for high energy expenditure. This needs rethinking. This is what the project will do: generate more sustainable tools for these sectors, as well as tools to measure energy expenditure in the different processes .

Meanwhile, Coloma Ballester, the IMVA coordinator, says: the digital media industries, where huge volumes of digital data are processed and post-processed, will benefit from research and transfer and will incorporate artificial intelligence, automating their tools and processes, while raising the quality of digital content and at the same time increasing the speed of production and reducing the demand for energy and resources .

Emerald's lines of action for applying AI to the media and audiovisual sector

To this end, the project will work with various goals. On the one hand, it will promote the automation of production processes, to save time and resources for the media or producers of audiovisual or entertainment content. On the other, it will develop AI-based machine learning applications to facilitate data processing in the industries of the sector or improve and optimize the quality of digital and audiovisual content. Work will also be carried out to promote energy savings and efficiency in these sectors. To monitor the results, a test bench will be created to measure and calculate the energy used by sector media and industries.

What tools will Emerald develop to improve the sustainability of these sectors?

The project will develop algorithms, techniques and tools to automate and improve the sustainability of specific tasks of media and audiovisual production processes. Some of them will be used to improve and optimize video editing and post-production, for example with regard video matting or work to correct the light and colour of films or audiovisual products after their assembly, and to reduce the bandwidth and energy needed to store and process them in the cloud.

Other tools, related rather to the dissemination of journalistic information, will be used to extract data from specific news or to synthesize or modify its contents. In addition, low-power devices for content analysis and new, more sustainable tools for relaying news will be created.

Other technologies will be more linked to sports information. Techniques will be developed that apply AI to the analysis of sport-related information and audiovisual content, as well as improving computer vision techniques that allow estimating the position of people or targets in real time during live audiovisual broadcasts. Also, a web server format will be created that is capable of automatically enriching the so-called metadata (descriptors of other data) in digital and audiovisual content from the world of sport, as well as to improve streaming. Other tools that the project aims to develop are linked to the application of AI in the world of animation and in the field of virtual and augmented reality.

Emerald also plans to develop prototypes of the tools designed to conduct demonstrations, which may encourage their subsequent development by other companies in the sector and their future entry into the market.

The European partners of the Emerald interdisciplinary consortium

UPF, through the GTI research group (DTIC-UPF)

UPF is coordinating the consortium behind the Emerald project, through the GTI (Interactive Technologies Group) and the IMVA (Intelligent Multimodal Vision Analysis Group, formerly IPCV) both at the Department of Information and Communication Technologies (DTIC) of the University. The GTI focuses on the human aspects of technology, especially on encouraging and improving the social use of ICTs, and develops projects related to inter

LINK:	https://www.filmlight.ltd.uk/store/news_articles/upf-emerald-project/...
	See more stories from filmlight

More from FilmLight