Autonomous Vehicles
We teach cars how to see the world
Training data for autonomous vehicles
At Figure Eight, we work with Tier 1 and Tier 2 suppliers, OEMs, and more to create high-quality training data for autonomous vehicles and self-driving car projects. We specialize in image annotation, doing everything from simple image classification to bounding boxes to semantic segmentation. So however you’re approaching the problem, we can help.
Our platform powers:
Bounding Boxes
Bounding boxes are often used to identify the most important classes in a computer vision model and our platform quality controls and baked-in machine learning make sure those boxes are accurate.
Lines & Polygons
Lines and polygons are typically used to mark lane lines and other important road markers for self-driving algorithms. Our line tool can also be used to outline oddly shaped or occluded objects.
Pixel Labeling Semantic Segmentation (PLSS)
More and more, autonomous vehicle algorithms are relying on exacting pixel-level labeling. We support dozens of classes per image and have the capacity to label massive datasets quickly–and most importantly–accurately.
Machine Learning Assisted Video Object Tracking
Our video object tracking solution leverages both human and machine intelligence to annotate video up to 100 times faster than human-only approaches.
Learn More
Figure Eight Machine Learning assisted Video Object Tracking
Learn how we combined the best of machine and human intelligence to create video training data faster and more accurately.
Trusted by today’s leading brands
Walking through the field, taking tallies on how many weeds we hit, how many we missed, and our accuracy was in the high 90s on that test. We’re really excited about bringing this technology to the farmer and using AI to really modernize farming.
Learn how Home Depot uses machine learning and high quality data to make every customer experience, no matter how they’re shopping, a whole lot better.
When we changed to Figure Eight, within a few weeks, we saw that labeler accuracy go up to 88% and it stayed in the high 80s and 90s for us ever since, even across a large diversity of models. That’s been a really, really big win.
Resources
FeaturedBlog
Introducing Machine Learning Assisted Video Object Tracking
In just three years, a million minutes of video content will cross global IP networks every second. That’s over 100......
Read MoreBlog
Introducing Instance-Based Pixel Labeling
There are plenty of ways to annotate images for computer vision projects. At a high level, you can simply bucket......
Read MoreeBooks
eBook: What We Learned Labeling 1 Million Images
From photos of earth from space to cellular microscopy and everything in between, we’ve seen our fair of images come......
Read More