Professor Rita Noumeir' research project
An ÉTS researcher’s technology could save the lives of suicidal inmatesMonday, February 8, 2021
Suicide in prison is more frequent than in the general population: it varies from 36 to 55 suicides per 100,000 inmates, compared to 11.5 in the Canadian population*. Rita Noumeir, professor-researcher at the École de technologie supérieure (ÉTS), believes that intelligent vision technology would help penitentiary staff intervene before an inmate commits the irreparable.
“It goes without saying that suicidal individuals need the support of professionals,” she begins. “However, can technology help penitentiary staff save lives by helping them intervene before an inmate commits the irreparable? I believe it can!”
Professor Noumeir, who has been working in image and video analysis for several years, is pursuing an industrial collaboration aimed at developing a tool to detect destructive and self-inflicted behaviour in prisons.
This project, carried out in partnership with Aerosystems International, would also help detect other high-risk situations, such as a fall, an overdose, or an assault between inmates.
To do this, the research team is working with state-of-the-art cameras that detect depth using infrared. By combining the flow of these cameras with that of 2D cameras, the application would be able to detect the position of the human body’s joints in space and understand how the body is moving. “That’s where AI comes in,” says the professor. Artificial intelligence analyzes this data to classify the action. Is there an action? When did it start? What is the action? “Does the person fall?” she uses as an example.
Challenges to overcome
The main challenge of this machine learning method, however, remains the availability of data for learning. “We don't have specific data for our application,” explains Professor Noumeir. The researcher and her team must use all their knowledge to take advantage of the public data that exists and specialize the algorithms.
Once the application detects problematic behaviour, it will always be up to a human being to check the inmate’s health status. “What we are doing is helping the human being,” concludes Rita Noumeir.
About Rita Noumeir
Rita Noumeir, professor in the Department of Electrical Engineering, develops applications integrating artificial intelligence. She is interested in the processing and analysis of videos, images and physiological signals. Her research work focuses in particular on activity recognition from video sequences and on decision support to improve the quality of intensive care for critically ill patients.
Professor Noumeir is a member of several scientific committees and is highly sought after by national and international funding agencies to evaluate research proposals.
Since joining the ÉTS in 1994, Professor Noumeir has contributed to the publication of more than one hundred articles in high-level peer-reviewed international journals or presented at international conferences.
As a lead researcher of numerous research projects that have received major grants from the Canada Foundation for Innovation, the Natural Sciences and Engineering Research Council of Canada (NSERC) – Collaborative Research and Development, Prompt and Mitacs, Professor Noumeir has established close ties with the Centre hospitalier universitaire Sainte-Justine as well as with companies in the medical field and specialized in security.
Find out more:
Artificial Intelligence Assisting Medical Diagnosis (Substance ÉTS)
Communication ServiceAll the news