How an AI Tool Transitioned from Confections to Cancer Diagnosis

Of all the technologies that have been proposed and developed over the years, few have tickled our collective imaginations the way that the concept of artificial intelligence has. For this reason, many small and medium-sized business leaders may write off AI as science fiction, rather than a solution to their operational challenges. However, the technology is currently finding its footing in practical use—more than proving its worth and adaptability in the process.

To help illustrate this, let’s examine how one AI platform has proven its capabilities in very different use cases… literally transitioning from the bakery checkout line to cancer diagnostics.

What’s the Point of AI in a Bakery?

To answer this, we must look to Japan, where many restaurants and other similar establishments operate in much the same way as a cafeteria would. Moving around the establishment, a customer collects what they are interested in eating and brings them to the checkout line.

In Japan, it isn’t unusual for a bakery to have dozens of items in different varieties, each with their own price. Furthermore, until somewhat recently, it wasn’t unusual for their baked goods to be presented without any wrapping. As you might imagine, this once drastically slowed down the checkout process—cutting into the revenue that these bakeries could once take in.

However, with the help of a particular AI, the role of the cashier in the checkout process was made much simpler. Rather than counting out and identifying each item individually, the cashier simply instructs their customer to place their items on the counter, where an overhead scanner can register them. The AI then identifies the pastries and calculates the total price in seconds, expediting the checkout process and thereby increasing the sales that the bakery can make.

Developed by BRAIN Co., “BakeryScan” costs approximately $20,000 to implement and can be found in over 400 establishments in Japan.

Note: If the idea of an unwrapped baked good being sold over the past year makes you nervous, it is important that we point out that BakeryScan has been improved to be able to recognize different baked goods through protective wrappings.

Bakery and Beyond

Once BakeryScan took off, it received some media coverage. This coverage ultimately led to its expansion into what few would likely expect: cancer diagnostics.

After watching a segment about BakeryScan on television, a doctor at Kyoto’s Louis Pasteur Center for Medical Research was struck by how similar some baked goods looked to certain cancer cells viewed through a microscope. What if BakeryScan could be adapted to determine key differences in cancerous cells, rather than confections?

The doctor quickly reached out to BakeryScan’s developers, and the tool was repurposed into AI-Scan.

What AI-Scan Can Do

Using the same technologies that can differentiate between anpan and shu kurimu, AI-Scan’s specialized Cyto-AiSCAN technology can identify the differences that are used to diagnose cancerous urinary cells with over 99 percent accuracy—and that’s just the start. Today, AI-Scan is in use in various industries, helping to distinguish between different kinds of pills in pharmaceutical work while also helping to identify potentially problematic fasteners and bolts in mechanical engineering. Toyota has even put AI-Scan to use to design a better airbag for their automobiles.

Hopefully, this will help give you a taste of how AI can be adapted to various purposes, helping to streamline the processes that many industries rely on. It’s also valuable to point out that many solutions today likely incorporate some form of integrated machine learning or AI that can benefit your business much in the way that AI-Scan has contributed.

Accucom is here to help you identify the best approach for you to take where your business’ IT is concerned and help you enact and maintain it. Give our experts a call at (02) 8825-5555 to learn more about what we can help you accomplish.