Success Stories
Artificial Intelligence

Stream Technologies' machine-learning solution can find photographic data hidden to the naked eye

Don Murphy, COO of Stream Technologies, COO of Stream Technologies
August 2, 2019

Share this page

For Stream Technologies, setting up shop at the Advanced Technology Centre is a lot like coming home.

Two of Stream Technologies’ co-founders, Don and John Murphy have a long history at the centre – they also co-founded Shana Corporation, which joined the ATC in 1988, shortly after the centre opened. From its small beginnings, that company grew to more than 100 employees before the Murphy brothers and their co-founders sold the business in 2003.

Now the Murphys are back at ATC with their new venture. And while a lot has changed, many things have stayed the same. “Many of the same benefits are still here, which is great,” says Don Murphy, Stream’s COO. “There’s a lot of bright minds in this building that we can interact with, share ideas and share solutions that we’ve come up with for different problems. That has been very consistent and very helpful.”

Stream is, at its core, a software service. Using a combination of machine learning and spectroscopy, Stream users can upload spectroscopic images, determine what they’re looking for, and then the software’s artificial intelligence creates a program to identify what the user selects. Because spectroscopic data spans a much larger portion of the spectrum than is visible to the human eye, the software can identify differences that may be invisible to us. The potential practical applications are huge, spanning everything from identifying the nitrogen content of soil to differentiating between precancerous and benign moles.

There’s a lot of bright minds in this building that we can interact with
Don Murphy Stream’s COO

What makes Stream unique is that the automatic creation of software is much faster and more accurate than manually creating algorithms to do the same work. “Typically, those kinds of applications can take six months or a year to develop. We can build those kinds of applications in hours or days,” says Murphy.

While the software itself doesn’t require a technical background to use, hyperspectral cameras are expensive, typically well over $60,000. Stream is working with Canon to develop an API call that would automatically send images from a regular off-the-shelf Canon camera to Stream for analysis and predictions.

The company’s focus on software also means more staff. In its seven years, Stream has grown from its three co-founders to 10 staff. Murphy credits the City of Edmonton and ATC with making it easy to find and attract talented people. “Edmonton’s actually one of the top areas where we have a pool of people that are very well versed in machine learning, for us to draw on.”

The flexibility of the ATC has also been a benefit to the fast-growing company. “We’ve moved twice since we’ve come in here,” says Murphy. “We started over in a small space and then based on the flexible leasing arrangements they’ve got we’ve been able to take up bigger space and grow into it without having to incur all those costs at the beginning.”

That flexibility could come in handy as Stream moves forward. Murphy says the company is hoping to expand further, and is in the midst of creating a new line of devices for examining soil and plant matter.

This story was first posted on the Advanced Technology Centre website.

You may also like