Just-in-time Visual Data Campaigning
Avoid exhaustive data collection with targeted campaigns identifying hard-to-find items and unknown scenarios at the edge
THE PROBLEM
THE STORY
Just-in-time Visual Data Campaigning
From Berlin to Las Vegas
How we enabled Visual Intelligence in Las Vegas
We brought our perception software into cars driving over thousands miles
We went to see professional drivers across the city
We gave them Aplomado CAMs to install in their vehicles
In 5 days, we covered a lot of distance
6
professional drivers
4808
kilometers driven
~5 million
scenes analyzed at the edge
hundreds
of data campaigns successfully conducted
CES 2022
Our driving partners enabled us to bring visual intelligence to the road and run visual data campaigns from our booth at CES.
Live demonstration of the engine sourcing capabilities
Introducing our tool: Scene Xplorer
Even though we had only 6 devices in the field, we were able to source visual context data & imagery for highly complex and not pre-learned situations as well as landmarks.
Jorit Schmelzle
Co-Founder and CPO
TECHNOLOGY
We engaged with attendees to allow them to run data campaigns in our tool.
Visitors at our booth had a lot of interesting scenes requests.
Valentine provides us with a picture of a crowded pedestrian crossing scene.
When a crowded pedestrian crossing was detected by our perception software at the edge, the scene was automatically captured, processed, anonymized and transmitted from the vehicle in real-time.
The scenes detected were collected in libraries. For this specific request, we generated 300 results in less than 24hours!
Our Scene Xplorer was able to run successful data campaigns, collecting millions of images of defined scenes under many different conditions - during daylight or night with sun or rain and from multiple angles.
Since we were in Las Vegas we also challenged our Scene Xplorer to obtain scenes including some of the most famous landmarks of the city - like the Montgolfier Balloon or the Strat Tower.
Our technology allows you to filter visual sensor data for very specific scenarios – for example, a night drive in the rain with five vehicles and two bicycles present in a round-about. Capturing data in pre-defined conditions and receiving it just-in-time will save significant resources and enable faster results.
Virginia Turner Taborda
Head of Growth
In 24 hours, a campaign is able to find thousands of images.
