New fractal promises faster hospital imaging

A fresh approach to sampling will make for better patient experience, researchers say. Andrew Masterson reports.

A visual representation of the novel fractal discovered by Shekhar Chandra and colleagues.

A visual representation of the novel fractal discovered by Shekhar Chandra and colleagues.

Chandra, et al

A new class of fractals discovered by an Australian researcher could make the unnerving process of undergoing a magnetic resonance imaging (MRI) scan significantly shorter.

In a paper published in the journal IEEE Transactions on Image Processing, Shekhar Chandra from the University of Queensland and colleagues describe a novel fractal arising from a key mathematical function known as the discrete Fourier transform.

The researchers apply the fractal in a sampling protocol to create an imaging methodology they term “chaotic sensing”. The turbulent nature of the results obtained means that they are independent – that is, they can be easily isolated.

Existing MRI methods process results using different methods of fractal sampling, which take longer to produce detailed imagery. Chaos sensing, the researchers write, achieves results more rapidly “while remaining capable of recovering a theoretically exact representation of the image”.

In practical terms, deploying the new fractal in MRI machines means data can be gathered much faster compared to existing methods.

“We’ve demonstrated that we can use the repetitive property of the pattern to our advantage to reduce the measurements required for an MRI scan,” Chandra explains.

And that should result in doctors obtaining the information they need in less time, thus decreasing patient discomfort and reducing waiting lists for the procedure.

“I believe we can be more intelligent in the way we collect our measurements and data to greatly improve patient outcomes,” Chandra says.

“My research proves that this sensing methodology can eventually be applied to many areas of science, including astronomy, biomedical engineering and computer science.”

Such applications will still be some time coming, however. Chandra and his colleagues have achieved proof-of-concept, but reprogramming the world’s MRI machines is still a long way off.

“In theory we’ve shown this discovery will improve MRI technology, but more work and more funding is needed before it can become a reality,” he says.

Latest Stories
MoreMore Articles