•New
Share
In the framework of a CENIT research project in which TV3 is participating
A team at the UPC designs technology that automates the production of football match highlights
A team from the Department of Signal Theory and Communications at the Universitat Politècnica de Catalunya (UPC) has designed a technology for automated summaries of football matches, including highlights. The system, which was developed in the framework of a CENIT research project in which Televisió de Catalunya participated, automatically selects video fragments that are likely to be of interest as highlights of the match.
07/01/2015
The technology was designed to edit an entire match automatically and extract a sequence of key moments. Francesc Tarrés, principal investigator for the project, explains that the technology "divides the video content into different frames, which are rated according to the relevance of the low-level information that is extracted from the images and audio using mathematical algorithms”.
This information includes aspects such as ambient noise levels, including the clamour from fans and the referee blowing the whistle; camera movements; predominant colours in the images; people or faces that accumulate in a particular segment of the image; etc. A numerical mark is assigned to each of these pieces of information, which goes towards a final mark for each of the video fragments that is detected by the computer. These marks are interpreted as tell-tale signs that something interesting is taking place on the field. The system selects all of these moments to create a summary of the game's highlights.
In a test of five games, the technology was able to select 70% of the goals. As one of the researchers, Arnau Raventos, explains, "We still need to improve specific combinations of moments in a match for the computer to be able to detect that what is happening is significant and should be included as a highlight". The researcher adds, "It is still not possible to produce a complete automatic summary," although the tool may help journalists and programme editors to select the highlights of a match.
The project was coordinated by Francesc Tarrés and carried out within the framework of a project belonging to the National Strategic Consortia for Technical Research (CENIT) programme. Researchers at the UPC's Department of Signal Theory and Communication Arnau Raventos, Raul Torres and Luis Quijada participated. All of the researchers mentioned, including the coordinator, are from the Castelldefels School of Telecommunications and Aerospace Engineering (EETAC), except Luis Quijada, who is from the Barcelona School of Telecommunications Engineering (ETSETB).
This information includes aspects such as ambient noise levels, including the clamour from fans and the referee blowing the whistle; camera movements; predominant colours in the images; people or faces that accumulate in a particular segment of the image; etc. A numerical mark is assigned to each of these pieces of information, which goes towards a final mark for each of the video fragments that is detected by the computer. These marks are interpreted as tell-tale signs that something interesting is taking place on the field. The system selects all of these moments to create a summary of the game's highlights.
In a test of five games, the technology was able to select 70% of the goals. As one of the researchers, Arnau Raventos, explains, "We still need to improve specific combinations of moments in a match for the computer to be able to detect that what is happening is significant and should be included as a highlight". The researcher adds, "It is still not possible to produce a complete automatic summary," although the tool may help journalists and programme editors to select the highlights of a match.
The project was coordinated by Francesc Tarrés and carried out within the framework of a project belonging to the National Strategic Consortia for Technical Research (CENIT) programme. Researchers at the UPC's Department of Signal Theory and Communication Arnau Raventos, Raul Torres and Luis Quijada participated. All of the researchers mentioned, including the coordinator, are from the Castelldefels School of Telecommunications and Aerospace Engineering (EETAC), except Luis Quijada, who is from the Barcelona School of Telecommunications Engineering (ETSETB).
Follow us on Twitter
