How Computer Vision is Used by AMP Robotics, Arturo, Zivid, Anduril, and Onfido: Case Studies

Many see computer vision as a next phase of artificial intelligence, where a machine can conjure habits, thoughts, and emotions, just like humans.

Today, computer vision has transcended the traditional use cases of spatial recognition and found a role in solving some of the fundamental enterprise value chain issues.

See below how five companies benefited from using computer vision in various sectors: defense, insurance tech, cybersecurity, waste recycling, and 3D vision systems:

1. Anduril

Seattle-based Anduril uses a combination of computer vision and AI to secure defense equipment for the frontline. Lattice, Anduril’s in-house infrastructure uses computer vision to build inspection systems and border surveillance guides, connecting ground radar with aerial jets and live sensors to detect and blend data into moving pictures.

“We founded Anduril because we believe there is value in Silicon Valley technology companies partnering with the Department of Defense,” says Brian Schimpf, co-founder and CEO, Anduril.

Anduril’s Lattice was also selected by the U.S. Air Force under Advanced Battle Management System, where it helped take down a mock missile using its computer vision algorithm. They have also built products for Ghost4 sUAS, Sentry Tower, and Anvil sUAS.

Industry: Defense technology

Computer vision solutions: Lattice

Outcomes:

• Spatial recognition to detect moving weapons, aerial systems, drones, and UAVs
• Surveillance system for U.S.-Mexico border and some other U.S. bases
• Creating a full-fledged war operating system with minimal human supervision

2. Arturo

Arturo’s USP lies in using a computer vision system to detect property casualties and determine insurance claims. The AI-powered sensors detect aerial, satellite, and ground-level images to verify damages and predict property data, thus minimizing the need for human intervention.

Currently, Arturo analyzes more than 71 types of property data and has completed 35 million inspections in Oceania and the North American region.

“Arturo’s vision is to revolutionize the way businesses interact with properties by providing greater transparency into a building’s physical makeup, and we believe we’re only scratching the surface in terms of the level of value we can bring to our customers,” says John-Isaac Clark, CEO, Arturo.

Industry: Insurance technology

Computer vision product: Arturo’s InsureTech

Outcomes:

• Simplifying underwriting, quoting, and claim settlements for both businesses and housing customers
• Multi-source imagery supporting 3D visualization of properties
• Detect climate impacts on properties and craft mitigation strategies
• Settle historical claims and determine the habitability of property within five seconds
• Confidence scores and predictive property analysis for maintenance and repairs
• API-enabled feedback loops

3. Zivid

Zivid creates 3D cameras that can give robots human-perfect vision. The enterprise uses computer vision to translate an input device into vision software for humanoids, cobots, and other industrial/household electronics. Zivid, with its recently launched Zivid Two product, boasts HD resolution and up to 55 μm imaging precision.

“Limitations like resolution, artifacts, and trueness errors in today’s vision systems restrict their potential in pick-and-place applications,” says Thomas Embla Bonnerud, CEO, Zivid.

“With the Zivid Two 3D camera, we’ve taken a giant leap forward in reducing these constraints. Using Zivid Two combined with deep learning, AI, and object detection algorithms, pick-and-place robots can recognize more objects, plan better grasps, and place more reliably. It all comes down to this — if your system can see more, it can do more.”

Industry: 3D vision system

Computer vision product: Zivid Two and Zivid One+ line

Outcomes:

• Creating second-generation cameras to enrich quality and take image recognition to a 10x scale
• Reduce logistics and create a shorter product development life cycle
• 3D precision with applications in piece picking, welding, soldering, bin-picking, and machine-tending
• Hassle-free storage of shiny and absorptive products where high precision was previous difficult
• A capture time of 0.3 seconds makes Zivid Two fit for industrial use

4. AMP Robotics

Denver-based AMP Robotics was born out of the sheer necessity of reducing and recycling waste through an integrated AI and computer vision mechanism.

The platform deploys computer-visioned robots with improved pick rates, 80 per minute, and holistic monitoring; automated waste segregation; easy adaptation to landfills and waste plants; and higher durability across waste types, including metals, plastic, polychlorinated biphenyls (PCBs), cartons, cardboard, thin films, etc.

“AMP’s robots have quickly doubled the pick rates we were accustomed to, maintaining and even improving the purity of our bales, which we depend on to maximize prices with our end-market buyers,” says Michael Cunningham, CEO, Recyco.

AMP Robotics recently went global with its first waste plant in Europe, taking the number to 160 plants set globally.

Industry: Waste recycling

Computer vision product: AMP Neuron, AMP Insights, and AMP Cortex

Outcomes:

• Machine-visioned robots identify, segregate, and disintegrate solid waste at twice the average speed
• Sorting and listing of materials across 100 categories with precision and transporting them to separate decomposers via AI-enabled pickers
• AMP Insights to visualize processing, key commodities, per-hauling load, monitor material activity, and real-time performance
• Flexibility to automate stock keeping units (SKUs)

5. Onfido

The brainchild of Oxford University, Onfido helps businesses handle cyberattacks through stringent identity verification protocol.

The idea of using computer vision in e-commerce stem from the aggravated attacks on financial services while the world migrated to online mediums, especially after the COVID-19 pandemic. Today, Onfido conducts KYC, secures financial transactions, and authenticates consumer backgrounds through computer vision and facial recognition.

“The current identity verification process is a big issue for underbanked and unbanked people. It’s also an insecure process: almost two trillion dollars yearly are laundered due to identity fraud. This is what Onfido is looking to solve,” says Husayn Kassai, CEO, Onfido.

Today, more than 1,500 global enterprises, including UberEATS, use Onfido to secure their platforms while ensuring smooth service standards for consumers.

Industry: Cybersecurity and identity authentication

Computer vision product: Onfido

Outcomes:

• Verifying a user identity without a need for physical visit — front camera selfies are the key
• Reducing human intervention in the form of clerical checks by comparing the selfie and user photoID through an AI-enabled software
• Automatic document validation when submitted and regular cross-checks with law enforcement, thus saving government resources