"Analysing facial gestures with artificial intelligence may help diagnose Parkinson's disease and dementia"

 Àgata Lapedriza

Àgata Lapedriza: "We research artificial intelligence to benefit humanity"

Rubén Permuy
Àgata Lapedriza, lead researcher of the Scene Understanding and Artificial Intelligence LAB group


Àgata Lapedriza is the lead researcher of the Scene Understanding and Artificial Intelligence LAB (SUNAI) group, which is associated with the UOC's eHealth Center and Faculty of Computer Science, Multimedia and Telecommunications. She has also carried out research at the prestigious Massachusetts Institute of Technology (MIT) on, among other subjects, the development of automatic systems for scene recognition based on deep learning, and she has been a visiting professor at Google.

Your group researches computer vision. What does it consist of?

Computer vision is the area of artificial intelligence (AI) that focuses on making it possible for machines to see, by which I mean understand the visual world. Seeing is very easy for us: we simply open our eyes. For a machine it is much more complicated. A machine captures images with cameras but then it needs AI to allow it to understand what is in them. It is this type of AI that we develop in our research group.

A key factor of your field is artificial intelligence. Do you think it will start to become more present in everyday life in the future?

Undoubtedly. In fact, it is now much more present than it was ten years ago. Every time you search for something on the internet or use your smartphone's virtual assistant you are using artificial intelligence.

Many airports have automatic customs systems that take your photo and compare it with the one in your passport to verify your identity, and a lot of gyms have access control systems that verify who you are based on your fingerprints. All this works thanks to artificial intelligence.

Some people are concerned that technological developments may lead to more social control and less privacy. What would you say to people with such fears?

I would tell them that researchers, companies, research centres and universities are extremely sensitive to this issue. Training is given on the impact of AI on privacy, there are ethics committees at research centres, as well as at companies and universities that conduct research, which ensure that the necessary ethical principles are upheld. There are also more and more regulations in place to make sure that AI is used only to benefit people.

I would advise them to read up on the measures that are taken to develop AI that is safe and positive for society, and also on the consent we all give in relation to the processing of our data.

I would also recommend not losing sight of the positive aspects that AI provides. For example, it is playing a very important role in medicine, and will become increasingly useful in the discovery of new drugs, for the early detection of diseases, for the design of new treatments and for personalized medicine.

You have been at the UOC as a member of faculty and researcher for more than 12 years. How do IT researchers and teachers manage to keep up with the pace of technological change?

We have to make an effort to stay up to date, taking into account the speed with which progress is being made in all areas. Conducting research, though, is fundamental if we are to offer cutting-edge education to our students.

The SUNAI research group is working on scene and facial recognition, even analysing emotions, with the use of technology. Could you give us some examples of projects in which you and your colleagues have participated? For example, you have stated that it can be applied to driving.

We currently have a lot of projects on the go. We are working on social robotics, applying our research to improving communication between social robots and people. This research is important for the development of care robots to, for example, have them help and provide company for elderly people who live alone. We are also working on assisted driving, developing new technologies to make the driver's experience at the wheel more positive and to make driving safer and less stressful.

We are also working on processing and classifying medical images, such as those of the retina, with a view to developing new disease recognition systems. We also have a lot of experience in the analysis of facial gestures, which is applicable to the early detection of diseases like Parkinson's or signs of dementia, given that in these cases the way the affected person moves their face is different to people who do not have these diseases. The differences may be very subtle and difficult to see, but a machine that analyses the minutest details of their movements may be capable of identifying them.

We have also worked on scene imaging, on the recognition of spaces. And, in fact, recently, we have been working on the recognition of natural disasters and incidents in images, such as hurricanes, storms, erupting volcanoes, fires and landslides, with the aim of detecting images on social media depicting events that may require assistance. This research is inspired by the needs communicated by the agencies that provide a humanitarian response to these types of phenomena, given that they have found that social media can be used to raise the alarm: when these types of events happen, social media are immediately flooded with images showing what is going on. The problem is these images currently have to be recognized manually, and this is a slow and expensive process. The aim of this project is to try to automate these processes to respond and provide humanitarian aid more quickly and efficiently.

Your research group has joined the eHealth Center, the UOC centre that specializes in digital health. What relationship is there between what you do and health? At the centre you have made reference to computer vision for health.

A large part of our activity is fundamental research into computer vision, that is, research into how to equip machines with the capacity to see. The ability to see is useful in many scopes, and health is one of them. I referred earlier to our social robotics projects, where the aim is to help people in lots of different ways.

For instance, one of our projects studies how to design robots that can help with emotional well-being, giving you positive psychology information and helping you do exercises to improve your emotional well-being and help you to learn techniques to better manage your emotions when necessary.

We have also worked with Sant Joan de Déu Barcelona Hospital and other organizations, such as MIT, Hyundai and Affectiva, on the reduction of stress in children aged three to ten who need surgery or invasive procedures. What we are doing is creating a motorized car equipped with the same technology we are developing in our assisted driving project. The idea is for children to travel from the waiting room to the operating theatre in this smart car. During the journey the car will interact with them, distract them and help them do exercises to relax. Additionally, the analysis of facial gesticulation may make it possible to diagnose diseases like Alzheimer's and Parkinson's earlier and monitor the evolution of sufferers. And, evidently, the work we have done on processing and classifying medical images is directly linked to health.

You have also been conducting research at MIT for a number of years. How would rate your experience in that American institution?

MIT has given me a lot. First, it gave me the opportunity to live abroad and see other ways of working, which is always an enriching experience. And MIT, and Boston in general, is a magnificent place. It is full of people doing cutting-edge research in a wide range of fields, and this makes it a very interesting, motivating and inspiring ecosystem. We currently have a number of projects under way in collaboration with MIT.

You have worked for Google. What did you do there and how did the chance to work there come about?

Google has a visiting faculty programme to fund visiting professors. It is one of the company's initiatives to establish and maintain communication and collaboration with universities. I was there for a year and the opportunity arose because we had research interests in common.

There is still a gender divide in technological studies and research. What would you say to a girl who is not sure whether to enter a field in which women are still in the minority?

I would tell her that if she likes science and technology to go for it, because they are fascinating. And, if she has any questions or doubts, to contact someone who can mentor her and offer her support.


The UOC's research and innovation (R&I) is helping overcome pressing challenges faced by global societies in the 21st century, by studying interactions between technology and human & social sciences with a specific focus on the network society, e-learning and e-health. Over 500 researchers and 51 research groups work among the University's seven faculties and two research centres: the Internet Interdisciplinary Institute (IN3) and the eHealth Center (eHC).

The United Nations' 2030 Agenda for Sustainable Development and open knowledge serve as strategic pillars for the UOC's teaching, research and innovation. More information: #UOC25years