We’re taking this show to Brussels! Next week, Mighty AI is a proud sponsor of AutoSens, a best-in-class summit for autonomous vehicles perception technology. While we’re there, our Chief Product Officer and Co-Founder Matt Shobe will join an elite lineup of speakers discussing the challenges that face today’s vision system engineers. (Seriously, check out this amazing agenda.) Shobe’s session will cover: • Enabling autonomous vehicles to recognize objects in context, no matter the weather, time of day or season • Reconsidering the image-labeling workflow and getting to the highest levels of ..read more

Getting training data for computer vision models that can detect and recognize objects in images is one thing. Training autonomous vehicles to see everything in the scene is quite another. What does an oncoming vehicle look like on a sunny summer day? How about on a rainy road at night? How do you accurately label objects that aren’t simple to define—like gaps in trees with sky in between? What about snow on the ground: Do you tag the snow, the ground, or both? In a webinar this week, Mighty AI Chief Product Officer Matt Shobe explored these challenges and shared how computer vision teams can ..read more

Pedestrian detection is one of the thorniest challenges that faces researchers building recognition models for self-driving cars. When you’re in “control” of a two-ton hunk of metal and glass moving at 75 miles per hour down a twisting road, a sharp understanding of everything on the road is safety-critical. That means computer vision researchers need to train autonomous vehicles to see, recognize, and understand the many different ways people look, move, and interact with the world—and ultimately how to respond to an infinite set of possibilities. At CVPR 2017 last week in Honolulu, researche ..read more

Mighty AI helps companies get the datasets they need to train and scale their AI models. You probably already know getting this data isn’t easy. And much of that work is even tougher for the teams building computer vision systems for autonomous driving and navigation. It takes lots and lots of data to train vehicles to understand their complex environments. Like, petabytes-generated-weekly lots—and that rate is increasing. That’s why we’re especially excited to announce our brand-new open dataset for baselining autonomous driving data. Now, this is not a comprehensive dataset with aspirations ..read more

News in the autonomous driving world never slows down (pun definitely intended), so let’s get right to it. Here are the top headlines from April, May, and June 2017: The autonomous driving market could create a passenger economy worth US$7 trillion by 2050. Read. That makes sense, since at least 44 corporations are working on autonomous driving projects. Read. Who are the leaders and the laggards in this movement? The Information has the scoop. Read. BMW, Intel, and Mobileye partnered with Delphi and—as of this week—Continental on their previously announced autonomous driving initiative. Read ..read more