EUR
en
- es
- pt
- fr
- de
- it
Published on 2025-02-18
A team from the Valencian University Institute for Research in Artificial Intelligence (VRAIN) at the Polytechnic University of Valencia (UPV) will use artificial intelligence techniques to reconstruct the 80,000 most damaged photos from DANA, where the deterioration does not affect central elements such as faces, body parts, or essential details.
Specifically, the specialists will employ two servers that will work to automate and restore around 200-300 photos daily using artificial intelligence. These are photos that could not be digitized or restored within the Fine Arts Faculty’s Recovering Memories project because their edges and frames are damaged, but their deterioration does not affect a central element such as faces.
It is estimated that out of the 250,000 photos submitted by families for this project, around 80,000 are in this condition. Meanwhile, approximately 35,000 have already been manually restored, digitized, and returned to their families.
The VRAIN team at UPV, responsible for reconstruction using artificial intelligence techniques, consists of researchers Carlos Montserrat and César Ferri, along with three Master's students in Artificial Intelligence, Pattern Recognition, and Digital Imaging (Miarfid): Nikita Averitchev, José Daniel Galván, and Hugo Albert, who are voluntarily contributing to the recovery of this valuable historical testimony. They have also sought funding from various private and public entities to secure resources for this project.
"The volume of photos we are working with requires significant automation for this photographic restoration, something only artificial intelligence can enable due to the vast amount of data involved. It is a complex challenge because not all photos have the same type of damage, and each contains different elements that need restoration. But the important thing is that families get back clean photos, just as they were before DANA, and in which they can recognize themselves," explains Carlos Montserrat, a researcher at VRAIN, UPV.
He adds, "This is why we are dedicating hours to this project and seeking funding to make it a reality."
A "FASTER AND MORE EFFECTIVE" PROJECT WITH AI
The Recovering Memories project is coordinated by the Fine Arts Faculty at UPV to restore the photographs and albums affected by the DANA storm on October 29.
After months of work, more than 46 students from the Fine Arts Faculty at UPV, under the supervision of professors Pedro Vicente, Pilar Soriano, and Esther Nebot, have taken a step forward to "accelerate the recovery of photographs and safeguard memory through technology and AI, because the difference between giving families a fully restored photo or a stained, damaged one is huge," explain the coordinators.
To date, 35,000 images have been manually restored, but "it is necessary to intervene digitally and with artificial intelligence to achieve a faster and more effective project so that those affected can recover their memories as soon as possible. Identifying areas for improvement, optimizing processes, and finding the best AI-driven solution, while ensuring that faces and recognizable body elements remain untouched," concludes Pedro Vicente.
COMMENTS
No customer comments for the moment.