Vision is a dominant sense used to interpret the world. Computers now have the same capability. Computer vision enables a machine or computer to obtain data and information from images or multi-dimensional data sources to transform them into valuable insights. Recent advances include facial recognition, gesture analysis, medical image analysis, pollution monitoring, and enhanced security and surveillance. It is revolutionizing industries such as manufacturing, healthcare, insurance, and more.
Given these new technologies, the world is buzzing with computer vision examples and how they are transforming the world we live in. Here's our breakdown of what the buzz is about:
Image recognition - Uses algorithms to recognize people, actions, objects, and images, typically before machine learning is used to define or understand the contents.
Cloud processing - Processes [images] via an outside server that has practically unlimited available resources, limited by upload and download internet bandwidth.
Application Program Interface (API) calls - Universally readable communication protocol of [image] data for easy integrations, such as into a customer-facing app, often via cloud processing.
Edge computing - Image recognition and analysis processed in or by the camera.
5 Leading Computer Vision Startups
In this article, we will introduce the leading computer vision startups and explain the diverse approach each has taken into the marketplace. Computer vision is a broad segment, with applications for everything from searching against product catalogs, to automating security, to analyzing production environments, to sorting food production, to reading human behavior. So, let us dig in.
Pilot.ai
Location: Palo Alto, CA
Employees: 30
Differentiator: End-to-end Powerhouse, both Edge and Cloud
Pilot.ai has quietly earned massive contracts with some of the biggest names in camera and video hardware, as well as becoming a preferred supplier to government. This Silicon Valley startup's list of customers is as impressive as it is confidential - for good reason!
Computer Vision Breakthrough Qualities:
- They are the intelligence behind major vision-related hardware, specifically brands that would prefer their customers not find out that their entire underlying vision intelligence was outsourced.
- Modern governments capture such a quantity of video and image data from satellites and crew-less aerial vehicles (drones) that it would be impossible to have trained human eyes monitoring even a tiny fraction. Thus they rely on Pilot.ai. Though we cannot disclose specific agencies or use cases, CEO & Founder Jonathan Su proudly asserts, "We do not do weapons," if that tells you anything.
- Their vision intelligence can even identify an active shooter, and enable automatic emergency alerts if plugged into a school or business' security cameras.
- Depending on the customer's specific need, Pilot.ai computer vision processes with the device (edge), in the cloud, or some combination of both. Unlike most computer vision companies, Pilot.ai even works with the customer to create a custom user interface and integrations, resulting in a real end-to-end vision intelligence solution. Though the most expensive of the providers, in this instance, you get what you pay for.
Shazura
Location: San Francisco, CA & Madrid, Spain
Employees: 12
Differentiator: Image "Fingerprints"
"We understand computer vision, but this is magic." -David Hettena, RocketSpace
Shazura is the brainchild of Sira Pérez de la Coba, a computer science engineer with 20 years of vision artificial intelligence (AI) experience, who leads a team in her native Spain. It has become one of the most powerful computer vision algorithms in the world. How? By having the courage to throw away the industry standard of labeled image data and starting over. Shazura, instead, captures an image's fingerprint - the aspects that differentiate it as unique - which cannot be obtained with simple text labels. Using this technique, Shazura can often learn effectively from just one product photo.
Computer Vision Breakthrough Qualities:
- RocketSpace initially selected Shazura as a top-tier technology for a Fortune 1000 corporate client as part of a corporate innovation services engagement. The C-suite executives' jaws hit the floor when Shazura proudly flipped to a slide explaining they hacked - gained access without specific approval - into the company's billion-dollar e-commerce website, scraped and indexed 30,000 images, and created an exponentially better ‘similar product recommendation engine,' all in about 17 hours. Their future technology partnership was approved the same day.
Nyris
Location: Berlin, Germany
Employees: 20
Differentiator: a Multi-pronged approach
Nyris, standing for ‘Neural Iris,' is a top European computer vision startup. Rather than being based around a single specific algorithm, Nyris takes the approach of layering on varying analysis technologies in sequence, and as needed, until the desired outcome is achieved.
Computer Vision Breakthrough Qualities:
- Nyris can analyze and return results in under one second and can report out the industry standard JSON object file with the level of detail you specify. The Nyris algorithm improves itself with each new image analyzed, so expect accuracy to develop and eventually surpass humans.
- With backing from the Google Startup Program in 2016, and now accepted as one of just ten European members in Microsoft ScaleUp, Nyris has achieved serious recognition.
20 Billion Neurons
Location: Berlin, Germany
Employees: 25
Differentiator: Understanding human physical expression
20 Billion Neurons (20BN) is using cameras and computer vision to understand human expressions and gestures- similar to how the human eye and brain would. Their computer vision algorithm can identify subtle and specific human actions, such as whether someone is staring at you, drinking a bottle of water, covering something with a piece of paper, and much more. Their performance benchmark is to be at least as accurate as human intuition. Imagine a computer being able to understand sign language!
Computer Vision Breakthrough Qualities:
With 20BN's technology, companies can:
- Collect actionable data on how customers physically interact with a retail layout and with each piece of merchandise, and reference that against sales
- Gain insights on how workers perform their tasks in a factory and identify factors that maximize efficiency and safety
- Eliminate the use of a mouse and touchscreen, and, instead, select and type in the air
- Use an app that enhances awareness and betters habits
EVK
Location: Graz, Austria
Employees: 20
Differentiator: Hyper spectral vision for autonomous robotic controls
Some startups emerge through a different journey than the typical path of Seed to Series A. In this case, EVK started as a hyper spectral imaging business. Upon transitioning to the family's next generation, CTO Matthias Kerschhaggl, and CEO Matjaz Novak have transitioned the company into, by all indications, a proper startup business. In addition to producing innovative hyper spectral camera products, they also create the intelligence to infer what the cameras are seeing.
Computer Vision Breakthrough Qualities:
- Hyper spectral cameras see beyond the visible light spectrum of red, green, and blue, and detect bands invisible to the human eye that allow certain materials to become highly visible. Water, dirt, plants, other organic matter, and even types of plastics become highly visible and differentiated when seen through these advanced sensors.
- EVK is the vision intelligence behind many waste recycling conveyor lines enabling the immediate identification of certain types of plastics for autonomous sorting. They are also deploying sorting and quality control solutions in agriculture by identifying imperfections in fruits and vegetables not easily visible to the human eye. This is just the beginning for EVK as the ability to capture and make decisions on hyper spectral image data is tremendous.
Increase Your Business Efficiency With Computer Vision Technologies
There are many ways computer vision can add tremendous value:
- Taking frustrating, repetitive tasks and automating them, freeing up workers to add value elsewhere
- Making data visible so that informed decisions are made
- Improving accuracy over-time due to self-learning algorithms
- Creating the ability to watch all things at all times compared to humans that can only focus on one point at a time
- Combining with artificial intelligence to infer context
Could computer vision enhance or even transform your business? RocketSpace's experienced Corporate Innovation Services Team can help you identify potential solutions, and help you find the perfect emerging technology partner to accelerate your company.
