Written By
Bianca de Loryn
College/Division
College of Science and Engineering
Publish Date
27 October 2023
Related Study Areas
Becoming an engineer
Master of Engineering Alumni Raiyan Talkhani thinks that by using artificial intelligence researchers can better focus on their projects and health professionals have more time for their patients. What he loves best about his job, however, is that he can work from his adopted hometown of Cairns.
Growing up in Saudi Arabia, Raiyan Talkhani was surrounded by industrious expats. “My Dad was a business development manager for a power company,” he says. Raiyan himself did not want to go into business. Rather, he wanted to become an engineer.
“For me, engineering is like solving puzzles. You take components from your data, you analyse them, and you identify whether or not the outcomes are what you were looking for, and whether it's working or not. That is what engineering is to me,” he says.
After finishing his Bachelor of Engineering in Malaysia, Raiyan decided to enrol at JCU in Cairns to study a Master of Engineering (Professional). “JCU allowed me to major in Data Science and the Internet of Things (IoT), which I was really interested in. When I applied in 2019, everybody in the industry thought that IoT was going to be the next big thing,” he says.
Mapping roads by using artificial intelligence
Raiyan says it took him some time to decide what kind of research project he wanted to do for his Master in Engineering, especially as his interests had moved on to artificial intelligence (AI) in late 2020.
“I was hell-bent on doing something AI-related because I knew that was what smart systems were going to be about,” he says. He adds that he decided to do a research project in cooperation with Jayden Engert, a JCU PhD candidate in Environmental Management.
“A lot of rural roads get built, but they're not added to official maps,” Raiyan says. “My Master's research project was an AI that can identify rural and remote roads using satellite images.”
Before they began the joint project, Jayden had drawn the roads by hand. Jayden gave the satellite images to Raiyan, whose task was to find and train an AI model that could identify whether the structure found was a rural road. Looking for specific items in images is called ‘semantic segmentation’, as it labels segments within an image as ‘road’ or ‘trees’.
“I ran countless experiments on a number of different models at the time, and then I picked the best four for my thesis,” Raiyan says. “At the time, these models were considered state-of-the-art, and I modified them to meet my needs.” Raiyan says he used Google Colab for this, which is a free service that allows running and testing computer code written in Python in one’s internet browser.
Training an AI to identify the content of an image is not as easy as it sounds, Raiyan says. “The problems we had were the edge pixels. Where does the road end and where does the tree line start?”
Finally, Raiyan’s AI was able to identify a road in a satellite image with an accuracy of around 70 per cent, which qualifies as good accuracy in the machine learning world.
Raiyan says the biggest difficulty in such a project is not the programming, but the reality that roads are being built continually, making some images outdated quickly, depending on the building activity in a certain area.
Finding a home in the JCU Ideas Lab
Since graduating in 2021, Raiyan has had plenty of opportunities to work remotely and collaborate with other researchers around Australia. Being based in Cairns, and with access to JCU’s innovation hub, the Ideas Lab, Raiyan was motivated to start his own business, Dronician Tech. Raiyan’s business focuses primarily on helping researchers commercialise their research projects, in a similar way as he had helped Jayden before.
The researchers Raiyan works with are usually funded by research grants, and for some of these grants, Raiyan says, the location of the staff may have an influence on how easily a project can obtain funding. “Working with collaborators from a regional or remote area may be a plus when applying for funding. Being based in Cairns, this gives me more brownie points,” Raiyan says and laughs.
An app for healthy smiles
One of Raiyan’s latest projects is an app that aims to connect dentists with patients from rural and remote areas. With Raiyan’s app, patients will be able to go to a nearby medical centre where five different photographs are taken and uploaded to the app.
The images will then be shared with the dentists so they can discuss potential treatments with the patient from their office in the city. “The next time a dentist comes to that town, they can go straight to treating the patient,” Raiyan says.
He estimates that by the end of this year, the app will be finished, and that it will help rural and remote patients to have improved access to vital oral health care.
Tips for future engineering students
Raiyan, who has recently become a co-director of the Cairns chapter of the QLD AI hub, says that graduates like to joke that studying engineering is difficult. “But the thing is, it's not as challenging as people say it is. Most of the AI systems are already built, so all you're doing is taking advantage of something that is already available,” he says.
"You're going to learn how these systems work and how you're going to use them for your specific project. It’s fun, and you also get to meet of a lot of scientists and researchers. They take pride in their work, and they are really lovely to work with.”