Synthetic data for better vision.
Today's state of the art deep learning computer vision algorithms require vast amounts of data to be collected, labeled, and cleaned -- this is expensive and can take months of work. At SBX we take a novel approach. For each of our vision systems we spin up a virtual world, train the algorithms in the simulation, and test them on real world data. By leveraging tools from VFX and video games we can achieve results that match and outperform the state of the art, in a fraction of the time -- take a look at our system in action!
The founding team draws on experience from companies like Uber ATG, Kindred AI, Yelp and Wish.