Yukon high school students from Faro have developed an artificial intelligence-powered tool to detect skin cancer early and hope it will earn top honours at the 2025 Canada-Wide Science Fair in Fredericton, N.B., from May 31 to June 7.
Olive Passmore and Lilou Lefebvre saw health-care challenges in the remote Yukon community of Faro, such as a lack of specialized equipment, not enough staff and small facilities, so they created a tool that lets people take photos of skin spots and checks the images for signs of cancer to help improve access to care in underserved areas.
Passmore and Lefebvre turned heads at the Yukon Stikine Regional Science Fair on March 26, presenting their project at Yukon University in Whitehorse. Chosen from about 400 finalists across the country, their work earned the pair a spot at the 2025 Canada-Wide Science Fair.
"I would say, even if we don't win anything, I think it's about the experience and making a change," Lefebvre said.
The students envision nurses at rural clinics using the app to photograph suspicious moles, with artifical intelligence analyzing images to flag high-risk cases. The duo said they hope this triage system could shorten the six-month wait to see a specialist, helping patients receive faster diagnoses and potentially life-saving treatment.
"The first thing for how we thought it would be used was when someone went to the nursing station and had concerns about a mole, the nurses would use our app that's integrated into the healthcare system and take a dermacopic image, and depending if it's cancerous or non-cancerous, you'll get prioritized on the wait list to see specialists, because, as I said, it can take up six months, which by then the cancer may be fatal," Passmore said.
To train their artificial intelligence model, the Faro students used nearly 19,000 images of skin lesions from an international database. They then created a smaller, more balanced dataset to boost accuracy, especially for hard-to-photograph areas, helping the model perform better in real-world conditions, they claimed.
The students built a simple app that works without internet. Users can upload or take a photo of a mole and the app runs it through the AI model. The result appears on screen, helping people in remote areas get faster assessments.
The Faro students told the News the project was mostly self-taught. Without access to local dermatologists for guidance, the students turned to online courses and open-source tools for coding. Their parents helped with design and presentation, while one father offered guidance on programming in Python, a coding language used to build the application.
"Our parents, they're both teachers, supported us, especially within like the graphic designing overboard and like bringing our board together," Lefebvre said.
"My dad helped us a little bit with the coding part, since he just knows a little bit about Python coding," Passmore said.
The 2025 Canada-Wide Science Fair awards ceremony will take place on June 5.
Contact Jake Howarth at jake.howarth@yukon-news.com
