ASX-listed weighing technology company Shekel Brainweigh and UK-based software provider Edgify have agreed to collaborate on the development of a system that would allow self-checkout machines to recognise fresh produce.
The plan is to use Edgify's approach to training AI systems at the edge rather than centrally. This, the companies say, will allow self-checkouts to be self-training.
Edgify's software is said to be able to achieve 99% accuracy, compared with the 50 to 60% accuracy of current systems.
Shekel and Edgify have yet to execute a binding agreement, but are already talking to major retailers around the world, according to officials.
"By combining our market-leading weighing technology with Edgify's software ability for deep learning training, we are in a fantastic position to present an enhanced self-checkout technology to the market to allow for a much more improved shopping experience," said Shekel Brainweigh CEO, Yoram Ben Porat.
"Theft at self-checkouts is also a major problem for retailers, with many incidents of consumers scanning through cheaper fruit and vegetables – something as simple as scanning an avocado as a brown onion – in order to avoid paying higher costs. With this new technology, we believe we can reduce shrinkage for retailers to help ensure they are running profitable operations," added Ben Porat.
Edgify strategy director Lucia Mancisidor said "This partnership with Shekel presents an opportunity to unlock the potential of using an edge device to enable ongoing learning for visual recognition in self-checkouts.
"With vast amounts of continuously changing data needed to train machines for visual recognition, this solution is yet to be recognised at a retail level. By using the self-checkout systems as a learning platform, we can avoid standard issues inherent with using the cloud – making operations for retailers more efficient and economical."