In just 22 years, DermNet has gone from an idea to a global website with plans for machine learning.
“DermNet is now the world’s most frequented site for skin diseases with more than 17 million medical professionals and their patients visiting every year,” says Amanda Oakley, Adjunct Associate Professor at the University of Auckland.
She founded DermNet in 1996, taking it from the early days of websites through to a site that is mobile-response and regarded as being at the forefront of internet dermatology.
In 2017, DermNet won a Ministry of Health Clinicians Challenge for a skin disease recognition tool. “We want to use Artificial Intelligence (AI) software to identify images of skin diseases through pattern recognition leading to quicker, easier and more accurate diagnosing,” she says.
Through this award, Amanda met Diana Siew who leads the external relationship of partnership for the MedTech Centre of Research Excellence. That has widened DermNet’s scope.
“Now, we’re exploring machine learning. Our collection holds hundreds of thousands of images, New Zealand has few dermatologists and AI may help screen out unnecessary referrals and help people look after themselves better,” says Associate Professor Oakley.
Anyone with a large photo library will understand the scope of creating order, so she brought PhD student Xiaoming Wang on board as an intern.
“Machine learning exists but we still need the human component to set up the library – from cross-indexing to removing personal information, adding unique codes, grouping images, location on the body and so on,” says Associate Professor Oakley.
Xiaoming was attracted to an internship for several reasons.
“DermNet has a history of providing world-class dermatological information, combined with their willingness to stay on top of the health technology sector by stepping into AI research, and creation of state-of-the-art apps for disease diagnosis.”
At DermNet, Xiaoming got the opportunity to dive into challenging tasks.
“I was leading my own project to organise an incredibly large amount of data. I was learning and extending my skills in image processing and AI. And there was the rare chance to work very closely with esteemed health professionals such as Dr. Amanda Oakley. I also gained an appreciation of the complex issues surrounding the health information sector.”
The internship benefited both Xiaoming and DermNet.
“I now have a chance to work long-term in world-class projects in digital dermatology and health, and my work with DermNet has given them to access large data-processing techniques to greatly boost their capabilities.”
Xiaoming has some advice for future interns.
“Learn as much as you can from the wider industry you’re interning in. The rarest gems are the chances to observe and participate in discussions with people from the industry, giving you an idea of how it operates. If nothing else, you’re going to be better equipped in interacting with these people in the future.”
DermNet is fully focused on the future with plans to evaluate machine learning and launch an online shop of images by the end of this year.
By Prue Scott