Posted on 12/10/2022 10:30:41 AM PST by libh8er
Amazon built a computer vision algorithm from scratch to identify products without barcodes to help robots ship products to you faster.
Robots may be the future, but robotic arms are apparently no good at using an old and steadfast form of technology: the barcode. Barcodes can be hard to find and might be affixed to oddly shaped products, Amazon said in a press release Friday, something robots can't troubleshoot very well.
As a result, the company says it has a plan to kill the barcode.
Using pictures of items in Amazon warehouses and training a computer model, the e-commerce giant has developed a camera system that can monitor items flowing one-by-one down conveyor belts to make sure they match their images. Eventually, Amazon's AI experts and roboticists want to combine the technology with robots that identify items while picking them up and turning them around.
"Solving this problem, so robots can pick up items and process them without needing to find and scan a barcode, is fundamental," said Nontas Antonakos, an applied science manager in Amazon's computer vision group in Berlin. "It will help us get packages to customers more quickly and accurately."
The system, called multi-modal identification, isn't going to fully replace barcodes soon. It's currently in use in facilities in Barcelona, Spain, and Hamburg, Germany, according to Amazon. Still, the company says it's already speeding up the time it takes to process packages there. The technology will be shared across Amazon's businesses, so it's possible you could one day see a version of it at a Whole Foods or another Amazon-owned chain with in-person stores.
The problem that the system eliminates -- incorrect items coming down the line to be sent to customers -- doesn't happen too often, Amazon says. But even infrequent mistakes add up to significant slowdowns when considering just how many items a single warehouse processes in one day.
Amazon's AI experts had to start by building up a library of images of products, something the company hadn't had a reason to create prior to this project. The images themselves as well as data about the products' dimensions fed the earliest versions of the algorithm, and the cameras continually capture new images of items to train the model with.
The algorithm's accuracy rate was between 75% and 80% when first used, which Amazon considered a promising start. The company says the accuracy is now at 99%. The system faced an initial hiccup when it failed to catch color differences. During a Prime Day promotion, the system couldn't distinguish between two different colors of Echo Dots. The only difference between the packages was a small dot that was either blue or gray. With some retooling, the identification system can now assign confidence scores to its ratings that only flag items its very sure are incorrect.
Amazon's AI team says it will be a challenge to fine-tune the multi-modal identification system to assess products that are being handled by people, which is why the ultimate goal is to have robots handle them instead.
More like “Amazon wants to SUPPLEMENT the barcode”.
If the AI can’t figure out bar codes there may yet be some hope that we will not have to worry about Skynet coming on line in the near future.
In the year 5555
Your arms are hanging limp at your sides
Your legs got nothing to do
Some machine’s doing that for you
Artificial intelligence is a combination of simultaneous nonlinear equations (AKA Neural Networks) that can be trained to recognize patterns, integrated with intricate “if/then” programming to perform primitive, repetitive tasks.
It is certainly artificial, but it is certainly not “intelligent.”
Correct. Won’t be very useful at a UPS hub where most everything looks like a brown brick.
“ Correct. Won’t be very useful at a UPS hub where most everything looks like a brown brick.”
****************************************************
I’m always amazed when I get an Amazon delivery of something like a bundle of six toothpaste tubes in a two foot by three foot box. Still some fine tuning needed in their Operations.
Amazon wants to eliminate the workers ...Robots are cheaper..Just have to have people who can fix them
Be fine by me if we can eliminate one more space-wasting thing on packaging.
Problem is, it won’t. All we’ve done is add those weird QR boxes, endless safety info and languages, and nutrition info.
The packaging has gotten real expensive just from the size needed to fit all the items.
Yeah. The local Mucky D’s went to human free kiosk in store ordering. You can NOT order a sausage McMuffin with egg. You can only order one as part of a combo. I walked out.
Then we read the article and discover that Amazon isn’t trying to replace the barcode. Instead, it wants to examine the physical shape of the object to verify if it is the right one being shipped.
As usual, a lying headline gets us to read the article, and by that time some people have forgotten what the headline said.
Amazon, I doubt you will take my advice, but here it is: during this scan to determine if the object is the right one, scan the barcode AND examine the image of the object. You’ll have a very low error rate if you do that.
Well, it should be easy for AI to keep track of a product’s point of origin: “China”…”China”…”China”…
We opened doors by thinking
We went to sleep by dialing “O”
We drove to work by proxy
I plugged my wife in, just for show
AI can figure out the barcode just fine. They’re running into a problem with FINDING where the barcode is. And Amazon is figuring out that the level of object recognition they’d have to program into the AI to figure out where the barcode is would just let the AI know what the object is without bothering to find the barcode. If the AI can figure out “this is object X and therefore the barcode is here” you don’t actually need it to do anything after X, the AI already knows what it’s got, forget the barcode.
Some products have packaging that is identical except for an S, M, or L for size or an added 'X' to suggest a premium version of a product.
I don't think you are right. Think about different sized items that all come in one size of box. Shoes for example. The AI knows it is a shoebox, but the size, style and color of the shoes inside is on the bar code.
Yeah but it’s not just about the box size. There’s colors pictures and text on the box. When a shoe store guy goes to the back room to get you the style and size you requested he ain’t using the barcode. He’s using the little square that has a picture of the shoe, the color (usually both in a square of color and with words), and size (usually in multiple measurement systems).
And as the article said, not all stuff is in nice neat 6 sided boxes. Some stuff is irregular. And even regular stuff who knows where they put the barcode. Anybody who’s actually watched a grocery store clerk, or done self checkout, knows it’s pretty random. Some companies like it on top where it can easily be scanned coming out of the shipping box. Some on the bottom where it’s easy for the clerk. Some like it on the front... I have no idea why. Some like it on the back. Heck I’ve even seen them on the side, which I’m pretty sure is just to be annoying. The level of teaching you have to do to a robot so it knows where Post puts their barcodes vs Betty Crocker vs Kellogg it’s already going to have know what the product is to know where the barcode is. And again, if the robot already figured out what the product is, scanning the barcode is just wasted time.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.