7 Startups That Bring Vision AI To The Edge

The vision AI industry is growing rapidly. There are more than a hundred startups in this segment, with more coming out of stealth mode regularly. These startups are classified as vision AI, machine vision, or computer vision AI. Regardless of the classification, they all basically do the same thing – apply machine learning algorithms to manipulate, monitor, and analyze video, images, and objects in any environment.

The things some of these startups do is straight out of science fiction, think of The Machine in Persons of Interest. This vision AI technology is not a few years away, it’s here now and in use. The technology is being used in cameras, drones, robots, consumer devices, industrial facilities, sports venues, retail, auto industry, medical labs, and so on. There are hundreds of use cases publicly available.

One startup provides a fleet of drones that can monitor any kind of landscape or environment, in order to help identify buildings, objects, and people, something that would be very useful to first responders in natural disaster situations. Another startup collected geospatial data on visitor foot traffic to Popeyes restaurants, helping identify a sales trend. Here are seven cool startups we’ve come across during our research efforts.

1. AEye

AEye co-founder Luis Dussan spent more than a decade working for Northrop Grumman, Lockheed Martin, and NASA, designing advanced systems in the areas of defense, reconnaissance, and deep space. While working there, he realized that self-driving vehicles faced similar problems to those in the defense industry. He left corporate America to found AEye, a startup that has developed an AI system for vehicles that is able to see, classify, and respond to its environment in real-time. The product is called iDAR,  and it helps vehicles think like a robot, but perceive like a human. The system components include 1) independent sensors that collect data 2) 2D camera that uses a homegrown innovation called Dynamic Vixels, which is a new type of sensor and 3) solid-state micro-optomechanical (MOEMS).

  • Company: AEye Inc.
  • Founded: 2013
  • HQ: Pleasanton, CA
  • # of Employees: 49
  • Funding: $59.1M
  • Founders: Luis Dussan (CEO) and Barry Behnken (SVP Engineering)
  • Products: Advanced vision hardware, software, and algorithms
  • Industry: Self-driving vehicles

2. Airobotics

The startup provides automated industrial drones solutions to large industrial facilities such as mining companies, oil & gas, seaports, and factories. The pilotless drones monitor onsite activity and provide data collection, data processing, and real-time analysis to users. The system can be used in surveying and mapping, inspection, tailing dams, stockpile management, haul road, security, and emergency response. In addition, the company has launched a developer program that enables customers to create new applications for its AI-driven data platform.

  • Company: Airobotics
  • Founded: 2014
  • HQ: Scottsdale, AZ
  • # of Employees: 209
  • Funding: $108.1M
  • Founders: Ran Krauss (CEO) and Meir Kliner (VP of R&D)
  • Products: Pilotless drone solutions
  • Industries: Mining, seaports, oil & gas, and others

3. Applitools

This startup provides a commercial-grade, visual AI-based cloud engine that helps developers shorten release cycles for mobile apps, web apps, and cloud-native apps. The system, powered by AI-computer vision algorithms is able to automatically test the look and feel of the user experience for any application, looking at UI elements such as color, shape, size, object position, and so on. The algorithm acts like a human eye and brain, analyzing every minute detail,  reporting on elements that once only the human eye could do. The SDK supports all the popular automation frameworks, CI systems, and issue trackers. The backend servers support the storage of a massive amount of screenshots for processing.

  • Company: Applitools
  • Founded: 2015
  • HQ: San Mateo, CA
  • # of Employees: 103
  • Funding: $41.8M
  • Founders: Gil Sever (CEO) and Adam Carmi (CTO)
  • Products: AI-based cloud engine that helps engineers in test automation, DevOps, and developers validate visual aspects of mobile apps, web apps, and native apps
  • Industry: Banking, retail, insurance, Pharma, software, online, and more
  • Customers: Salesforce, Mastercard, SAP, Sony, Intuit, Dow Jones, Addidas…

4. Nauto

The four-year-old startup has been very successful, providing an AI-powered behavior learning platform to more than 350 fleet companies. The in-vehicle system helps prevents collisions. It uses a proprietary technology called in-Cabin Alerts. The in-Cabin system monitors driver behavior and notifies distracted drivers of pending danger, helping avoid collisions, near misses, and traffic violations. If an accident does occur, the system provides the collision video to streamline claims processing. Also, the system identifies high-risk drivers and alerts their employers.

  • Company: Nauto
  • Founded: 2015
  • HQ: Palo Alto
  • # of Employees: 186
  • Funding: $173.9M
  • Founders: Dr. Stefan Heck (CEO)
  • Products: AI-powered driver behavior learning platform
  • Industry: Commercial fleets safety
  • Customers: 350+ fleets including Hertz, LeasePlan, Sprint, Zenken…

5. Neurala

Neurala has developed a platform powered by deep learning neural networks that makes devices like cameras, drones, robos, toys,  and self-driving cars intelligent. The main product is called Brain Builder, and it uses a proprietary technology called Lifelong-DNN that enables customers to create image classification models with ease. Users can build prototypes, tag data, performs training, then roll out models into production with ease. The system provides an in-depth analysis of the devices in use. One of the benefits of integrating Neurala into a product, the time spent on data preparation is reduced by more than 90%. In the data science field, scientists typically spend 80% of their time in data preparation. Nvidia, a major hardware player for the AI industry has partnered with Neurala to provide an edge-based deep learning solution.

  • Company: Neurala
  • Founded: 2006
  • HQ: Boston, MA
  • # of Employees: 49
  • Funding: $15M
  • Founders: Dr. Massimiliano Versace (CEO), Dr. Heather Ames (COO), Dr. Anatoli Gorchet (CTO)
  • Products: Vision AI
  • Industry: Across several industries
  • Customers: Motorola, Squadrone System, DARPA…

6. Orbital Insight

The startup provides an AI-based Geospatial system that uses satellite imagery, UAV, synthetic aperture radar, AIS, and geolocation signals to ingest, interpret, and analyze large volumes of data. Thereafter, it contextualize the raw data to provide actionable intelligence to its users so they can mitigate risk. The system can do interesting things like access foot traffic in a region to identify key trends, aggregate billions of  cell phone geolocation pings anonymously, visualize customer foot traffic at trade shows, detect land-use changes over time, detect the number of cars and operating vehicles in a given area, classify land use, monitor construction areas, and the list goes on. In one interesting project, the startup was able to monitor foot traffic to PoPeyes restaurants and identify a key sales trends.

  • Company: Orbital Insight
  • Founded: 2013
  • HQ: Palo Alto, CA
  • # of Employees: 128
  • Funding: $78.7M
  • Founders: James Crawford (CEO)
  • Products: Startup provides Geospatial solutions uses sources such as satellites and UAV to capture and process data.
  • Industries: Cross-industry. Working with Fortune 500, hedge funds, humanitarian orgs, and more.

7. Shield AI

Shield AI is a robotics company that builds products for the national security sector and first responders. They enable robots to see, reason about, and search for an environment and its surroundings. One of their creations is Hivemind, a drone that is able to search buildings, create maps of the area, and stream videos of it. The company is also able to help in coordinated exploration. In this type of deployment, a group of robots acts in unison to develop a structural model of an environment by navigating through it. Sensors on the robot (drones) see the environment, learn its structure and conditions, then share the data with each other and its operators. This type of service is helpful in disaster recovery incidents and military use cases.

  • Company: Shield AI
  • Founded: 2015
  • HQ: San Diego, CA
  • # of Employees: 105
  • Funding: $48.1M
  • Founders: Ryan Tseng (CEO), Brandon Tseng (COO), and Andrew Reiter (Technical Fellow)
  • Products: Robots that can see, reason, and search for things.
  • Industry: National security sector and first responders.