Watch for These 8 AI Startups Doing Computer Vision
We cover topics about artificial intelligence quite often on Nanalyze because it promises to be one of the most disruptive technologies in our lifetime. Some experts called futurists believe that the Singularity—the moment in time when machines will become exponentially smarter and more capable than humans, turning us into biological Duracell batteries a la The Matrix—is only a few decades away. We’ll soon witness the first machine wars, as companies develop AI cyber security systems to battle machine-generated malware that can learn how to avoid detection. Computers are learning to code themselves as well as develop software. We can expect that all sorts of jobs will go away in the future, even (or perhaps especially) white-collar occupations.
Of course, we were promised jetpacks and flying cars decades ago, so it’s sometimes hard to imagine how all of these AI advances will eventually play out. Many seem abstract and far removed from our everyday lives. But there is one advancement in AI that we see all the time: computer vision. Our definition of computer vision is a field of research for designing machines with the ability to process, understand and use visual data just as humans do. Probably one of the best examples out there is the ability of Google to do image recognition. The company recently showed off its AI abilities by indexing about four million photographs from Life Magazine without any human intervention. A website allows you to search by tags, such as bikini or cycling.
We’ve covered a number of computer vision companies in the past (here and here, for example), but we’ve been keeping our eye out for some more interesting startups in this space because it’s one of the key technologies that will drive automation, from driverless cars to robotic construction workers.
Let’s take a closer look at eight computer vision startups that have taken in new funding within the last year.
China In Your Face
We recently did a deep dive on one of China’s biggest AI companies, SenseTime, which is a leader in facial recognition, especially when it comes to government surveillance. But SenseTime isn’t the only well-funded facial recognition startup in China, which is rapidly accelerating its AI technology across the board, including into AI chips. There are a few other Chinese heavyweights in this space: Megvii, CloudWalk Technology, and Yitu Technology.
Founded in 2011, Beijing-based Megvii has raised a ton of money since we first featured it about a year ago. Total disclosed investments now total $607 million, following a $460 million mega Series C last November, making it yet another Chinese company valued at $1 billion or more. Investors include Ant Financial and Foxconn, which manufactures iPhone components. It has developed a cloud-based facial recognition platform for mobile and gaming applications. Forbes reported that the company’s FaceID platform has identified and indexed 290 million faces worldwide. The company also boasts more than 85 percent market share in security software for China’s financial industry, according to Forbes.
CloudWalk Technology is based in Guangzhou, which is rapidly turning itself into one of the tech capitals of the world. In fact, the city gave the AI startup a $301 million grant to set up an AI computer vision center in the city. It recently closed a $75 million Series B, bringing total funding to about $376 million. The company’s products reportedly include facial recognition terminals, facial scanning door entry and something called infrared binoculars scanning machines. China Money Network reported that its customers include more than 100 financial institutions, while its big-data facial recognition system is used in police departments in 23 provinces.
Founded in 2012, Shanghai-based Yitu Technology has taken in about $60 million in disclosed funding last year, with Sequoia Capital on the roster of investors. Its Dragonfly Eye Intelligent Security System is another platform used to promote the state surveillance operatus catch bad guys. For example, police in the city of Xiamen have used facial recognition technology to arrest 18 suspects and crack down on 60 burglaries, robbery, and pickpocketing cases. The platform also does cars, searching for vehicles by image and analyzing fake license plates. It’s not all about law and order: Yitu also applies its computer vision tech in healthcare for medical record management and even clinical diagnoses from X-rays and other imagery.
Update 07/18/18: Yitu Technology bagged $100 million in additional funding in July of this year, just one month after receiving $200 million. That brings total funding to just over $355 million.
Turn the Other Cheek
By now you may be feeling a bit paranoid and considering wearing a Guy Fawkes mask wherever you go in public. Don’t worry. Three veterans of the Israeli Defense Forces have your back—and your face. Tel Aviv-based D-ID raised $4 million in January, a year after the startup went live. Its pedigree includes the Y Combinator accelerator program known for picking long-term winners. It’s proprietary technology somehow “scrambles” images so that computer vision algorithms aren’t able to identify faces but the altered images aren’t discernible to the human eye. The product is primarily intended to protect imagery stored by banks, governments, and other institutions from hackers.
The Self-Flying Camera
Here comes another nail in the coffin of the once-ubiquitous GoPro: A drone equipped with 13 cameras that flies autonomously using computer vision, following you on whatever daredevil stunt you can imagine. The S1 drone is a product of Skydio, a Silicon Valley startup that has raised $70 million in disclosed funding, including a $42 million Series B in February. It’s got some big names in its corner, including Andreessen Horowitz, Accel Partners, and Nvidia, which also provides some of the hardware behind its AI acrobatics. The Skydio R1 drone is hands-free flying via an app. Tap your human target and the drone will follow. Want to record your own death-defying hijinks? It will only set you back $2,499. We wonder how many the Chinese have on order.
Looking Good with Computer Vision
Not every company in China is using computer vision to keep tabs on its citizens. Founded in 2014, Malong Technologies in Shenzhen has raised about $46 million in disclosed funding, with about $35 million of that coming in a Series B last November. Its investors include names like Microsoft and Softbank, while its partners also include Nvidia and eBay. The startup has won a mantle shelf full of awards, but perhaps none more prestigious than WebVision, the Olympics of the computer vision world. Malong won the competition, which involves classifying 2.4 million images that have not been annotated or edited by humans into 1,000 categories, with a 94.78 percent recognition rate. It applies it computer vision technology, Product AI, in a number of ways, from searching on visual fashion accessories (not the first company to build AI into fashion) to automating photo tagging:
As you can see, the algorithms do a pretty good job, though we would argue that this “beautiful” shot is not “monstrous” or “poor” by any means.
More Than Meets the Eye
Founded in 2014, Paris-based Prophesee, formerly known as Chronocam until recently, took in $19 million in a Series B last month, bringing total disclosed funding to $37.3 million. Investors include Intel and Renault, among others. The startup was also listed on the CB Insights AI 100 list in 2017. Prophesee says its computer vision technology is based on neuromorphic engineering principles, meaning that its systems help machines mimic the human eye and brain. The bio-inspired tech can process 100,000 video frames per second for applications in robotics, industrial automation, and self-driving cars, among others:
The technology is rooted in advances that helped researchers discover the so-called God particle, the Higgs Boson, one of the Holy Grails of particle physics. Perhaps that explains the name change.
Facial Recognition for Cows
Founded in 2015, Dublin-based Cainthus suddenly became famous when Cargill announced it would make an undisclosed investment in the Irish computer vision startup. Cainthus focuses on facial recognition technology for cows. Yes, cows. The idea is that machines can monitor the animals remotely and send alerts when action is needed, such as when the animals need food or water. It can even detect if an animal is lame and needs medical attention. AgFunder News reported that farmers can earn up to $200 per animal per lactation cycle by optimizing livestock care. And that’s no bull.
As you can see, computer vision has many real-world applications. While quite a few seem to skew toward Big Brother surveillance, there are a number of creative commercial applications coming to market. And though some of the machine wars will be fought in cyberspace, no doubt the real battles (figuratively and otherwise) will be fought among companies for supremacy in automation, self-driving cars, and robotics, among other industries.
It’s also good to note that half of these startups are from China, which is aggressively competing in the AI arena. It has the cash on hand to outspend anyone at the moment. However, China’s vision of the future may not be one that most of us want to see.