Here’s why Amazon wants to get rid of the barcode

• Amazon has a plan to get rid of barcodes from its warehouses.
• The e-commerce giant is training robots that can identify items while picking them up and turning them around.
• Products in Amazon warehouses will still need barcodes as long as companies that make and ship them use them.

Robots may be the way of the future, but it turns out that robot arms can’t read a good old barcode. Barcodes can be hard to find and can be put on products with strange shapes, which robots can’t fix very well.

So, Amazon announced on Friday that it has a plan to get rid of barcodes.

Using pictures of items in Amazon warehouses to train a computer model, the e-commerce giant has made a camera system that can watch items as they move down conveyor belts to make sure they match their pictures. AI experts and roboticists at Amazon want to use this technology to make robots that can identify items while picking them up and turning them around.

Nontas Antonakos, an applied science manager in Amazon’s computer vision group in Berlin, said, “Fixing this problem is essential so robots can pick up items and process them without having to find and scan a barcode.” “It will help us get customers’ packages to them faster and more accurately.”

The multi-modal identification system won’t completely replace barcodes any time soon. Products in Amazon warehouses will need to have barcodes as long as the companies that make and ship them use the technology to identify and track stock.

Amazon said that its new system is already being used in facilities in Barcelona, Spain, and Hamburg, Germany, where it is speeding up the way packages are handled. The technology will be used by all of Amazon’s businesses, so you might see a version of it one day at Whole Foods or another store chain owned by Amazon.

Amazon has added computer vision to some of their other products. “Alexa, what am I holding?” is a question you can ask an Echo Show smart display to help you figure out what you’re holding.

The Show and Tell feature was made with people who have trouble seeing in mind. AI has also been added to camera and photo apps by companies that make smart phones and social media sites. For example, photos are automatically put into categories.

Amazon says that the problem that the system solves, which is wrong items being sent to customers, doesn’t happen very often. But even small mistakes can cause big delays when you think about how many items a warehouse processes in a single day.

Amazon’s AI experts had to start by putting together a library of pictures of products, which the company hadn’t done before. The first versions of the algorithm were trained with both images and information about the size of the products. The cameras continue to take pictures of items to train the model with.

When it was first used, the algorithm worked between 75% and 80% of the time, which Amazon thought was a good start. The company says that it now works 99% of the time. At first, the system had trouble because it couldn’t tell the difference between colours. During a Prime Day sale, the system couldn’t tell the difference between Echo Dots of two different colours.

The only thing that made one package different from another was a small dot that was either grey or blue. With some changes, the identification system can now give its ratings a confidence score so that it only flags things that it knows for sure are wrong.

Amazon’s AI team says it will be hard to fine-tune the multi-modal identification system to evaluate products that are being handled by people. This is why the ultimate goal is to have robots handle them instead.

Also Read

US lab achieves fusion milestone, fostering optimism for clean energy

US scientists made significant advancement in fusion energy on Tuesday. Researchers at...