“In the real world, zoom and enhance can only get us so far.” Those were the words uttered in the Castle television series, after the main character was disappointed with the quality of an enhanced image. This was probably a dig taken at other movies and shows which have long given the impression that there exists computer software capable of enhancing reflections and other low-quality images to yield detailed high-quality images. Unfortunately, this is not the case… not yet, at least.

Poor picture quality may be caused by factors such as compression (to reduce storage space), low resolution (loss in detail due to a low number of pixels composing an image), and wide-angle lenses (which capture a wide field of view at the expense of decreasing the object resolution). Such factors cause a loss of information that cannot be easily recovered, since this in turn increases the number of ways in which an image can be reconstructed.

An easy solution would be to use better cameras. However, these are more expensive to purchase, while higher-quality footage typically consumes more storage space. For larger buildings requiring a substantial number of cameras, costs can rise to the point where the installation of cameras may not be possible, compromising security. Moreover, there already exist numerous low-quality images and videos that may contain some useful information. Hence, their quality also needs to be improved somehow.

the rate of progress in this hot research field is such that, maybe someday, there will exist methods capable of coming close to the techniques portrayed on television

To this end, the Deep-FIR project has been researching how Artificial Intelligence (AI) can be applied to perform ‘super-resolution’, where algorithms are designed to essentially reverse the degradation processes and restore the original high-quality image by recognising patterns and deducing the most plausible underlying image.

The project has considered not only the improvement of synthetically degraded images (as is often done by existing methods), but also images afflicted with real-world degradations. While there still remain several challenges, the rate of progress in this hot research field is such that, maybe someday, there will exist methods capable of coming close to the techniques portrayed on television.

Project Deep-FIR (R&I-2017-002-T), composed of  John Abela, Kenneth Camilleri, Christian Galea,  Matthew Aquilina and Keith George Ciantar, was done at the University of Malta in collaboration with Ascent Software, and was financed by the Malta Council for Science and Technology (MCST), for and on behalf of the Foundation for Science and Technology, through the FUSION: R&I Technology Development Programme.

Christian Galea was a post-doctoral researcher within the Faculty of Information and Communication Technology at the University of Malta and works primarily in the fields of AI, machine learning and computer vision with an interest in their application to areas such as biometrics, forensics and autonomous driving.

Sound Bites

•        Scientists have created an AI system capable of generating artificial enzymes from scratch. In laboratory tests, some of these enzymes worked as well as those found in nature, even when their artificially-generated amino acid sequences diverged significantly from any known natural protein.

•        A temperature reconstruction from ice cores of the past 1,000 years reveals that today’s warming in central-north Greenland is surprisingly pronounced. The most recent decade surveyed in a study, the years 2001 to 2011, was the warmest in the past 1,000 years, and the region is now 1.5 °C warmer than during the 20th century, as researchers report. Using a set of ice cores unprecedented in length and quality, they reconstructed past temperatures in central-north Greenland, and melting rates of the ice sheet.

https://www.sciencedaily.com/releases/2023/01/230118111656.htm

For more soundbites, listen to Radio Mocha every Saturday at 7.30pm on Radju Malta and the following Monday at 9pm on Radju Malta 2 https://www.fb.com/RadioMochaMalta/.

DID YOU KNOW?

•        Apple’s latest transistors are only 25 atoms wide.

•        In Colombia, you can vote for ‘none of the above’ in elections.

•        About one person in 20 cannot visualise images in their head.

•        The city of Iquitos in Peru has a population of over 400,000 but is not reachable by road.

For more trivia, see: www.um.edu.mt/think.