CLOSE
iStock // Ekaterina Minaeva
iStock // Ekaterina Minaeva

Man Buys Two Metric Tons of LEGO Bricks; Sorts Them Via Machine Learning

iStock // Ekaterina Minaeva
iStock // Ekaterina Minaeva

Jacques Mattheij made a small, but awesome, mistake. He went on eBay one evening and bid on a bunch of bulk LEGO brick auctions, then went to sleep. Upon waking, he discovered that he was the high bidder on many, and was now the proud owner of two tons of LEGO bricks. (This is about 4400 pounds.) He wrote, "[L]esson 1: if you win almost all bids you are bidding too high."

Mattheij had noticed that bulk, unsorted bricks sell for something like €10/kilogram, whereas sets are roughly €40/kg and rare parts go for up to €100/kg. Much of the value of the bricks is in their sorting. If he could reduce the entropy of these bins of unsorted bricks, he could make a tidy profit. While many people do this work by hand, the problem is enormous—just the kind of challenge for a computer. Mattheij writes:

There are 38000+ shapes and there are 100+ possible shades of color (you can roughly tell how old someone is by asking them what lego colors they remember from their youth).

In the following months, Mattheij built a proof-of-concept sorting system using, of course, LEGO. He broke the problem down into a series of sub-problems (including "feeding LEGO reliably from a hopper is surprisingly hard," one of those facts of nature that will stymie even the best system design). After tinkering with the prototype at length, he expanded the system to a surprisingly complex system of conveyer belts (powered by a home treadmill), various pieces of cabinetry, and "copious quantities of crazy glue."

Here's a video showing the current system running at low speed:

The key part of the system was running the bricks past a camera paired with a computer running a neural net-based image classifier. That allows the computer (when sufficiently trained on brick images) to recognize bricks and thus categorize them by color, shape, or other parameters. Remember that as bricks pass by, they can be in any orientation, can be dirty, can even be stuck to other pieces. So having a flexible software system is key to recognizing—in a fraction of a second—what a given brick is, in order to sort it out. When a match is found, a jet of compressed air pops the piece off the conveyer belt and into a waiting bin.

After much experimentation, Mattheij rewrote the software (several times in fact) to accomplish a variety of basic tasks. At its core, the system takes images from a webcam and feeds them to a neural network to do the classification. Of course, the neural net needs to be "trained" by showing it lots of images, and telling it what those images represent. Mattheij's breakthrough was allowing the machine to effectively train itself, with guidance: Running pieces through allows the system to take its own photos, make a guess, and build on that guess. As long as Mattheij corrects the incorrect guesses, he ends up with a decent (and self-reinforcing) corpus of training data. As the machine continues running, it can rack up more training, allowing it to recognize a broad variety of pieces on the fly.

Here's another video, focusing on how the pieces move on conveyer belts (running at slow speed so puny humans can follow). You can also see the air jets in action:

In an email interview, Mattheij told Mental Floss that the system currently sorts LEGO bricks into more than 50 categories. It can also be run in a color-sorting mode to bin the parts across 12 color groups. (Thus at present you'd likely do a two-pass sort on the bricks: once for shape, then a separate pass for color.) He continues to refine the system, with a focus on making its recognition abilities faster. At some point down the line, he plans to make the software portion open source. You're on your own as far as building conveyer belts, bins, and so forth.

Check out Mattheij's writeup in two parts for more information. It starts with an overview of the story, followed up with a deep dive on the software. He's also tweeting about the project (among other things). And if you look around a bit, you'll find bulk LEGO brick auctions online—it's definitely a thing!

nextArticle.image_alt|e
REM-Fit
arrow
Live Smarter
Stop Your Snoring and Track Your Sleep With a Wi-Fi Smart Pillow
REM-Fit
REM-Fit

Everyone could use a better night's rest. The CDC says that only 66 percent of American adults get as much sleep as they should, so if you're spending plenty of time in bed but mostly tossing and turning (or trying to block out your partner's snores), it may be time to smarten up your sleep accessories. As TechCrunch reports, the ZEEQ Smart Pillow improves your sleeping schedule in a multitude of ways, whether you're looking to quiet your snores or need a soothing lullaby to rock you to sleep.

After a successful Kickstarter in 2016, the product is now on sale and ready to get you snoozing. If you're a snorer, the pillow has a microphone designed to listen to the sound of your snores and softly vibrate so that you shift positions to a quieter pose. Accelerometers in the pillow let the sleep tracker know how much you're moving around at night, allowing it to record your sleep stages. Then, you can hook the pillow up to your Amazon Echo or Google Home so that you can have your favorite smart assistant read out the pillow's analysis of your sleep quality and snoring levels the next morning.

The pillow is also equipped with eight different wireless speakers that turn it into an extra-personal musical experience. You can listen to soothing music while you fall asleep, either connecting the pillow to your Spotify or Apple Music account on your phone via Bluetooth or using the built-in relaxation programs. You can even use it to listen to podcasts without disturbing your partner. You can set a timer to turn the music off after a certain period so you don't wake up in the middle of the night still listening to Serial.

And when it's time to wake up, the pillow will analyze your movements to wake you during your lightest sleep stage, again keeping the noise of an alarm from disturbing your partner.

The downside? Suddenly your pillow is just another device with a battery that needs to charge. And forget about using it in a place without Wi-Fi.

The ZEEQ Smart Pillow currently costs $200.

[h/t TechCrunch]

nextArticle.image_alt|e
iStock
arrow
Design
Forget Horns: Some Trains in Japan Bark Like Dogs to Scare Away Deer
iStock
iStock

In Japan, growing deer populations are causing friction on the railways. The number of deer hit by trains each year is increasing, so the Railway Technical Research Institute has come up with a novel idea for curbing the problem, according to the BBC. Researchers there are using the sound of barking dogs to scare deer away from danger zones when trains are approaching, preventing train damage, delays, and of course, deer carnage.

It’s not your standard horn. In pilot tests, Japanese researchers have attached speakers that blare out a combination of sounds designed specifically to ward off deer. First, the recording captures the animals’ attention by playing a snorting sound that deer use as an “alarm call” to warn others of danger. Then, the sound of howling dogs drives the deer away from the tracks so the train can pass.

Before this initiative, the problem of deer congregating on train tracks seemed intractable. Despite the best efforts of railways, the animals aren’t deterred by ropes, barriers, flashing lights, or even lion feces meant to repel them. Kintetsu Railway has had some success with ultrasonic waves along its Osaka line, but many rail companies are still struggling to deal with the issue. Deer flock to railroad tracks for the iron filings that pile up on the rails, using the iron as a dietary supplement. (They have also been known to lick chain link fences.)

The new deer-deterring soundtrack is particularly useful because it's relatively low-tech and would be cheap to implement. Unlike the ultrasonic plan, it doesn’t have to be set up in a particular place or require a lot of new equipment. Played through a speaker on the train, it goes wherever the train goes, and can be deployed whenever necessary. One speaker on each train could do the job for a whole railway line.

The researchers found that the recordings they designed could reduce the number of deer sightings near the train tracks by as much as 45 percent during winter nights, which typically see the highest collision rates. According to the BBC, the noises will only be used in unpopulated areas, reducing the possibility that people living near the train tracks will have to endure the sounds of dogs howling every night for the rest of their lives.

Deer aren't the only animals that Japanese railways have sought to protect against the dangers of railroad tracks. In 2015, the Suma Aqualife Park and the West Japan Railway Company teamed up to create tunnels that could serve as safer rail crossings for the turtles that kept getting hit by trains.

[h/t BBC]

SECTIONS

arrow
LIVE SMARTER
More from mental floss studios