The concept of mobility is changing course to a service-oriented ecosystem. Ride hailing and sharing providers are multiplying, and scooter sharing solutions are flooding metropolitan hubs. Aside from electric vehicles and autonomous vehicle advances, our experience of getting from point A to point B is changing and the notion of smart mobility services is permeating into the corners of the automotive industry. Contributing to that change is Israeli startup UVeye by enabling smart mobility with deep learning and computer vision based vehicle inspection services.
UVeye provides high-end solutions for automatic external inspection of vehicles, using advanced technologies that include proprietary hardware combined with deep learning and computer vision algorithms.
The startup was founded by Amir Hever (CEO) and Ohad Hever (COO) in late 2016, originally targeting the homeland security sector to automate the human process of vehicle security inspections, traditionally conducted with mirrors angled at a vehicle’s undercarriage. Both Hever brothers are experienced tech executives with expertise in AI-first startups. And with global sales from the start, the startup’s technology and product-market-fit proved promising, prompting the brothers to focus and expand the offering to commercial applications too. Today, the startups clientele ranges from automotive OEM to car dealerships and fleet managers.
UVeye offers their solution in the form of three products: Helios, Atlas, and Artemis, covering the vehicle undercarriage, exterior, and tire inspection respectively. The startup combines software and hardware in computer vision to captures images of the car, while stationary and driving.
Their system utilizes multiple high-speed, high-resolution (up to 25 mega pixel) cameras, that generate a high resolution image of the vehicle’s undercarriage and exterior in real time and provides analysis within several seconds. With the generated image, their patent-pending image-processing algorithms identifies vehicle anomalies, like weapons and contraband, or faulty vehicle components, that are often symptoms of internal maintenance issues.
“A mechanic can now conduct their inspection on a computer without touching the vehicle” explained Hever. “We’re focusing on the automotive industry and providing the automatic solutions to the lifecycle of the vehicle from manufacturing to assembly, the dealership, and to the consumer… from 10,000 km to 100,000 km. And we’ve become the preferred standard for OEMs.”
The startup’s commercial offering is designed with classic computer vision algorithms, generally understanding beforehand what the proper appearance of most vehicle undercarriages should look like. The possibilities are less sparse. Whereas for threat detection, it involves identifying threats, known and unknown in appearance. “We developed a collection of AI algorithms that individually provides some information on the possibility of a threat in the car, such as convolutional neural networks for semantic representation, to detect all possible threats, even those that the system never encountered before” explained VP R&D, Yossi Shvartzman. The offering is available on premise and cloud.
UVeye’s undercarriage 3D imaging. Photo: UVeye.
They’ve scanned millions of car images to date and work on virtually every vehicle type. The startup lists automotive corporation clients, like Toyota Tsusho, and Skoda, as well as local bus fleet company Kavim. They total 80 employees today and are planning to expand their team by double over the next year, including 25 roles in their AI team (hiring).
“We want to enable smart mobility with our end-to-end inspection-as-a-service and set the standard for future vehicle inspection” stated Hever. “Our system enables car manufacturers to conduct inspection processes with more precision and speed, allowing them to increase volume.”
The startup is an alumni of the DriveTLV FastLane program, which enabled and assisted them with global and local partnerships, such as bus transportation company Kavim.