By Susanne Togeby, AAU Communication and Public Affairs. Translated by LeeAnn Iovanni, AAU Communication and Public Affairs. 
Photo: Hans Raven 

Walking very close to the harbour’s edge is abnormal. Taking a walk along the harbour is normal. Falling into the water is very abnormal. The artificial brain learns to distinguish between normal and abnormal, and over time it gets smarter and smarter. This is how Neelu Madan explains the model she developed through her PhD project at Aalborg University in collaboration with Milestone Systems, a company that deals with surveillance. 
 
"Implementing this system can benefit society because it can report abnormal behaviour. The challenge is that it will be close to impossible for a human to browse through such large amounts of data. With my research, we can automate this process," says Neelu Madan. 

The artificial brain can learn to sound the alarm in case of accidents 

An obvious place to use the model is at Danish ports where many drownings occur when it is dark and people are alone. This also applies to the waterfront in Aalborg which is extra dangerous because the harbour is so close to Jomfru Ane Gade where many people are under the influence of alcohol.  
 
With Neelu Madan's research, which combines artificial intelligence and surveillance, the artificial brain can learn to detect if a human is drowning – and sound the alarm.  The AI model is trained using large amounts of data from video surveillance that shows normal behaviour on the waterfront. The more video material the artificial intelligence is fed, the better it will be at detecting when something abnormal occurs. That way, the artificial brain can also make decisions in situations it has not experienced before. 
 
Neelu Madan's model is not only useful in drowning accidents, but also in all other places or situations where accidents can happen, such as nursing homes or hospitals where the artificial brain can use its experience with normal behaviour to assess if an accident occurs.  
 
According to Neelu Madan, manufacturing companies can also benefit from the model which can detect whether automated processes are not working as they should. That way, the company can detect errors that would otherwise be expensive.  
 
For CT and MRI scans in hospitals, Neelu Madan's research may be very important. The artificial intelligence can review the scans and immediately indicate if it sees anything other than normal organs. The AI can, for example, detect a tumour in the brain that the doctor would not be able to notice because humans cannot look through such large amounts of data.  

Neelu Madan receives Spar Nord Foundation's research award for her extraordinary research 

Neelu Madan's research is so extraordinary that the Spar Nord Foundation selected the young researcher to receive the research award in 2024. This took place at a ceremony on April 19 in Musikkens Hus for Aalborg University's 50th anniversary.  
 
The Spar Nord Foundation Research Award is given to a junior researcher who in the past year completed and had approved a particularly excellent PhD project. 
 
The prize comes with DKK 250,000 which Neelu Madan expects to use for further research in the medical field.