Inflated Auto Quotes: ProovStation Provides GPU-Driven AI Ratings

Inflated Auto Quotes: ProovStation Provides GPU-Driven AI Ratings

Vehicle ratings are pumped up with a GPU-accelerated AI overhaul.

ProovStation, a four-year-old startup based in Lyon, France, is embarking on the ambitious computer vision quest of automating vehicle inspection and repair estimates, aiming for ultra-high-resolution, computer-driven stations. AI in businesses around the world.

It recently launched three of its state-of-the-art vehicle inspection scanners at French retail giant Carrefour’s Montesson, Vénissieux and Aix-en-Provence sites. ProovStation drive-thru scanners are deployed in Carrefour car parks so drivers can pull over to experience the free service.

Self-service stations are designed for users to provide vehicle information and walk away with a value report and repair estimate in less than two minutes. It also allows drivers to get an offer from the dealer to buy their car as quickly as seconds, which is promising for consumers, as well as used car dealers and auctioneers.

There’s a lot at play in cameras and sensors, high-fidelity graphics, multiple damage detection models, and the models and analytics to turn damage detection into repair estimates and purchase offers.

“People often ask me how I managed to put so much AI into this, and I tell them it’s because I work with NVIDIA Inception,” said Gabriel Tissandier, general manager and product manager at ProovStation.

Leveraging NVIDIA GPUs and NVIDIA Metropolis SDKs allows ProovStation to digitize 5GB of image and sensor data per car and apply multiple vision AI detection models simultaneously, among other tasks.

ProovStation uses the NVIDIA DeepStream SDK to build its sophisticated vision AI pipeline and optimizes AI inference throughput using Triton Inference Server.

The setup allows ProovStation to perform inferences for fast vehicle scan times on this revolutionary industry-leading artificial intelligence application.

Driving progress: Bernard Groupe dealers

ProovStation deploys its stations in the blink of an eye. This was possible because founder Gabriel Tissandier partnered with an ideal ally in Cédric Bernard, whose family Groupe Bernard car dealerships and services first invested in 2017 to boost its own operations. .

Groupe Bernard collected massive amounts of image data from its own companies for the ProovStation prototypes. Bernard left the family business to join Tissandier as co-founder and CEO of the startup, and co-founder Anton Komyza joined them, and it’s been a wild ride of launches since.

ProovStation is a member of NVIDIA Inception, a program that accelerates cutting-edge startups with access to hardware and software platforms, technical training, and support from the AI ​​ecosystem.

“People often ask me how I managed to put so much AI into this, and I tell them it’s because I work with NVIDIA Inception,” said Tissandier, General Manager and Product Manager at ProovStation. .

Launch of AI stations in all markets

ProovStation has deployed 35 scanning stations so far and plans to double that number next year. It has launched its powerful AI-driven state-of-the-art stations in Europe and the United States.

Early adopters include Groupe Bernard, UK vehicle sales site BCA Marketplace, OK Mobility car rentals in Spain and Sixt car rentals in Germany. It also works with undisclosed US automakers and a major online vehicle seller.

The car rental company Sixt has set up an agency at Lyon Saint-Exupéry airport in order to facilitate the collection and return of cars.

“Sixt really wants to change the experience of renting a car,” said Tissandier.

Creating an “AI super factory” for damage datasets

ProovStation has developed data science expertise and a dedicated team to manage its many specialized datasets to meet the difficult challenge of damage detection.

“Going from a damage review to a damage estimate can sometimes be very tricky,” Tissandier said.

ProovStation has a team of 10 experts in its AI Super Factory dedicated to labeling data with its own specialized software. So far they have processed over 2 million images with tags, defining a taxonomy of over 100 damage types and over 100 part types.

“We knew we needed this level of precision to make it reliable and efficient for businesses. Image labeling is extremely important, especially to us, so we have invented ways to label specific damage,” he said.

Tissandier said data science team members and others are updated on AI with courses from the NVIDIA Deep Learning Institute.

Data Collection with NVIDIA Industrial Edge AI

ProovStation scans a vehicle with 10 different cameras in its station and takes 300 images – or 5 GB of data – to run on its detection models. NVIDIA GPUs activate ProovStation’s AI inference pipeline in 90 seconds to provide detection, damage assessment, localization, measurements and estimations. The wheels are scanned with an electromagnetic frequency device from the tire company Michelin for wear estimates. It all works on the NVIDIA edge AI system.

Using two NVIDIA GPUs in one station allows ProovStation to process all of this into high-resolution image analysis for improved accuracy. This data is also transferred to the cloud so that ProovStation’s data science team can use it for further training.

Cameras, lighting and positioning are big issues. Detection patterns can be disturbed by things like reflections on shiny cars. ProovStation uses a defectometry model, which allows it to perform detection while projecting lines onto vehicle surfaces, highlighting the points where problems appear in the lines.

It is a difficult problem to solve that leads to business opportunities.

“The whole auto industry inspects cars to provide services — to sell you new tires, to fix your car or your windshield, it always starts with an inspection,” Tissandier said.

Similar Posts

Leave a Reply

Your email address will not be published.