HR4RS stamp
Raul Garcia Hemonnet
Dan Casas is a young researcher from the Higher Technical School of Computer Engineering (ETSII) of the Rey Juan Carlos University who has been working in the field of digitization for more than 10 years.
Now, he has received one of the 58 'Leonardo' scholarships that the BBVA Foundation awards to innovative projects in different fields of science.
Support and recognition
For Casas, this fact means “both financial support and recognition of a career, since I have been in this field for almost more than 10 years. This aid (40.000 euros) will be used to hire equipment, buy technology, and be able to travel to conferences or 3D studios”.
Dan Casas is one of the 58 researchers from 9 fields of knowledge selected by the BBVA Foundation from a total of 1.200, “it is a joy and a source of pride. We always make proposals, sometimes with more success and others with less. It is important to always persist. This is the third time I've tried. Although it is a very competitive world, by insisting, something is achieved”, he explains.
Personalized digital avatars for all audiences
Specifically, thanks to this grant, Dan Casas will be able to continue his work in the line of research 'Personalization of 3D digital avatars with self-supervised machine learning'
Currently, he explains, “when we talk about digitizing humans, it is usually done with a 3D scanner or a semi-professional studio where you have many cameras. This digital avatar can be used in video games and movies. Also in the fashion industry since by having a reliable avatar we could try on virtual clothes”.
The young URJC researcher points out that “there are other methods based on Artificial Intelligence (AI). From seeing many examples of avatars and photos, computers learn to generate these avatars. AI needs many pairs of examples. When the computer has seen thousands of images, it learns to do the task”.
Dan Casas is working on a method that allows these computers not to need so much 'training', “the computer will learn to do the task without seeing many examples. The goal is to do it with a single photo. The machine would learn to extract the 3D information from a 2D image, trying combinations of relief until it matches the image it knows."
Using less images and data reduces the time of some very expensive processes. The objective of Dan Casas and his team is to reduce these costs to universalize access to personalized human avatars.
This would make it possible, for example, to make it a daily occurrence to go to the doctor with the virtual avatar, try on clothes, or talk to friends who are far away.
In the medium term, Dan Casas points out, “what will happen is that we will all have avatars of ourselves because it has many advantages. It will enable more immersive communication.” Or for example, in the field of Medicine, it will be used to prepare an operation and make "virtual consultations", indicates this young researcher, also for the fashion industry, since his team also works on the digitization of fabrics for achieve more realistic virtual fabrics.
Learn the behavior of human masses
Right now, in addition to this project, Dan Casas is immersed in an ambitious European initiative to digitize people at massive events. The project, called CrowDNA, is endowed with 400.000 euros and is part of an even larger one, endowed with 4.000.000 euros.
CrowDNA, in which important European institutions and companies participate, studies the behavior of the human masses with the aim of preventing risks and also designing safer spaces for holding massive events.