13 Computer Vision Companies You Need to See
Computer vision is the one artificial sense that we need to master first. If you think about the way you value your own senses, as much as you love tucking into a succulent piece of wagyu beef with your favorite Prince song playing in the background followed by a massage with some lovely incense wafting through the air, you’ll probably give up any one of those senses in favor of your sight. Without your sight, you’re helpless. Likewise, a computer or a robot can do an incredibly large amount of tasks if it has computer vision that allows it process what it sees as well as a human.
We talked before about a publicly traded company called Cognex Corporation (NASDAQ:CGNX) that builds computer vision hardware, but there is also a gold rush of software startups trying to attack all kinds of computer vision applications and some of these are just plain fascinating to read about. Let’s take a look at 13 computer vision startups that will give you some conversational fodder for that boring office holiday party you’re going to get dragged to next month.
|Company||Founded||Funding||Hood||Cool Stuff Computer Vision Can Do|
|Blippar||2011||$99 million||London England||Letting Your Favorite Brands Sell You More Isht|
|Captricity||2011||$52 million||Berkeley||Reading Your Horrible Handwriting|
|Affectiva||2009||$34 million||Massachusetts||Seeing What Makes You Laugh and Cry|
|Blue River Technology||2011||$30 million||Silicon Valley||Cops, Coming to Try and Scan Your Crops|
|uSens||2013||$27 million||San Jose, Cali||Giving You Artificial Limbs in Virtual Reality|
|Zebra Medical Vision||2014||$20 million||Israel||Making Sure You Don’t Die|
|Hover 3D||2011||$19 million||San Fran||Taking Your Home And Making it Virtual|
|Mint Solutions||2010||$12 million||Amsterdam||Assuring You That You’ve Taken All Your Drugs|
|Nanit||2015||$7 million||New Yawk||Monitoring Your Baby So You Don’t Have To|
|Kairos||2010||$4 million||Miami||Picking Out Your Beautiful Face From a Crowd|
|Pixoneye||2014||$4 million||London England||Looking Through Your Mobile Phone Pictures|
|Umbo CV||2014||$3 million||Taiwan||Straight Getting Some Criminal Profiling On|
|Smartspot||2015||$2 million||San Fran||Telling You How to Get Your Lift On Bro|
We covered these blokes before and we’re not taking the piss but their solution was a bit underwhelming. Is their business model really all about getting you to scan brand logos so they can give you some recipes or you can play some games? We weren’t blown away by our Heinz brand experience but the fact that this company is now a mythical unicorn (worth a billion dollars) means that they have a lot of potential. Just think about it. What if you had an app that helps you recognize anything you point it at? That’s where they are heading no doubt.
The first thing that you read when you visit Captrcity’s website is that “Captricity extracts and transforms data from handwritten and typed forms at 99.9+ percent accuracy“. That’s pretty much all that needs to be said, right? That has to be the best elevator pitch ever written, and they admit that their technology is “often first met with doubt and skepticism”. All you have to do is fax or email them some documents and they’ll transcribe them automagically. They make it so easy that you almost have to give it a go.
Just a few days ago Facebook bought a company like this and people speculated that soon you can simply like or dislike a post on Facebook by using your natural facial expressions. Talk about getting real. What’s even better is that you can actually let their computer vision AI agent observe you while watching a commercial and it will tell you how you reacted throughout. This is a key piece of technology that will let the robots of the future know when they’re getting on our nerves.
In the last two years, Blue River Technology has built a “see-and-spray” system that combines 3D computer vision, robotic sprayers that fire 40 times a second, and artificial intelligence to identify plants and act on them. Weeds are hit with herbicides and crops are fed fertilizer. This technology is being deployed in machines that eliminate 90 percent of the chemicals growers use today by replacing broadcast spraying with targeted controls. If you buy lettuce, there is a 1 in 10 chance they’ve cared for it.
The idea here is that when you are wearing virtual reality (VR) goggles, you can affix a piece of hardware to the goggles that will track the movement of your hands in real-time which is then relayed to the VR software. The idea is that as you hold your hands out in front of yourself in virtual reality, you will actually see your hands move as they are moving in the real world. The hardware tracks 22 points on your hands and fingers and can also be used for head tracking or to relay other elements of your environment into the virtual world creating an augmentation of real-world over virtual world.
In a previous article, we talked about how you should probably discourage your children from getting into radiology. Deep learning algorithms are now able to look at an x-ray and make a diagnosis with greater accuracy than humans. That’s precisely what Zebra Medical Vision is doing. They have developed “insights” that allow them to make predictions and diagnose diseases. For example, one of their insights can identify cardiovascular disorders and could save up to 500,000 lives. They plan to develop 100 more of these insights over the next 3 years. They have plenty of partners using their platform which provides access to millions of anonymized, indexed patient imaging studies, reports, and associated clinical information.
The ability to measure the surface area of a home just by taking some pictures with your smartphone would be most appreciated by building contractors who have to do this sort of thing all the time. This is the target market that will be expected to pay for this cloud-based solution that extrapolates loads of useful data from pictures taken with your smartphone. The other obvious use is for the homeowner who wants to see their 3D home rendered digitally and then play around with different paint colors, roofing tiles, sidings, etc. Below is a diagram of how this computer vision platform works.
Update 04/25/19: Hover has raised $25 million in Series C funding to continue growing and further developing their technology. This brings the company’s total funding to $67.3 million to date.
Mint Solutions has built a suite of products called Medeye designed to verify that a patient has received the correct administration of medication. With the Medeye scanner, a nurse can simply put all of the medications for a patient in a disposable tray, scan in seconds and verify they are correct. Unlike with barcodes, hospitals can verify all 5 rights – including the right dose – all in one go. According to Mint Solutions, 1 in 5 administrations of medication are wrong (wrong dose, wrong route, wrong patient, wrong time, or wrong medication). The computer vision comes into play by recognizing all the medication and then reconciling that with the patient’s electronic health record.
In a previous article, we talked about 5 baby monitoring companies that will keep an eye on your precious bundle of joy. One of these companies, Nanit, is best described as the “Tesla of baby monitoring devices”. The technology uses advanced computer vision and machine learning algorithms to analyze your baby’s sleep patterns and provide you with personalized tips and insights. It’s “the baby monitor that thinks” and they’re all sold out until January 2017.
Similar to Affective that we discussed earlier, Kairos is also working on detecting human emotions with computer vision (also called affective computing), but in addition to that, they are also working to understand the emotion, demographics, and characteristics of faces in video and photos. The Company talks about having the power to identify people in real-time and then cater your marketing message to them specifically. We’re not sure whether to be excited or afraid when we think about ads that know who we are when we simply look at them. They also let you test your video content against an audience of people whose emotions are being measured.
Pixoneye’s softer claim is that they “harness the full potential of image understanding on mobile devices to provide you with the ultimate key to capture the consumer’s ever-changing needs“. Sounds pretty cool right? Then the next sentence goes on to say “Our technology analyzes consumers’ personal photo galleries on their devices anonymously, without ever seeing or moving the photos, and creates digital profiles filled with clever data that helps you understand your audience’s personal life story with unprecedented precision“. Does anyone else find that to be really, really intrusive? This goes back to that whole data privacy thing we were talking about a few weeks ago.
There are a lot of security cameras in use that are “dumb” because they need a human to interpret what they are seeing. That’s all going to change with the Umbo CV “cloud-based” solution that uses machine learning to record what it sees. The more it is used, the better it gets at recognizing what is going on because the cloud-based algorithms are constantly learning. You can now use your security cameras to learn about your business and customers during the day and guard against intruders during the night. It can identify cars, bikes, or pretty much anything that moves around in much the same manner you or I do. Again, just a tad bit creepy how good machines are getting at monitoring what we’re doing.
Update 10/08/19: Umbo Computer Vision has raised $8 million in funding to push its sales efforts globally and double down on AI product development. This brings the company’s total funding to $17.6 million to date.
We wrote an article before about what happens when you use AI to pick startups to invest in. This is one of the startups the AI algorithm picked and it’s a cool smart “mirror” that watches while you pump iron and makes sure you’re not cheating on your rep counts. As you can see in the below example, it’s also using augmented reality to show you data in real-time. When you bench press your own bodyweight, it will congratulate when you log into “the cloud” to see how your workout went later that day. Can’t remember if it’s arms or legs day? Your smartphone will for sure. Lifting will never be the same again bro. You’re definitely going to want to buy tickets to this gun show.