Data centric AI

Machine learning (ML) and deep learning has proven to outperform traditional algorithms in many computer vision tasks. The next key challenges are the lack of both: the availability training data with accurate ground truth annotations and methods for capturing such data.

To solve the overarching data challenge in ML and AI we have developed disruptive methods that enable generation of high quality synthetic data for a wide range of applications - from autonomous vehicles to medical applications. We use state-of-the-art image synthesis for accurate simulation of sensors, and generative models to synthesize data with pixel-perfect annotations for ML training, testing, and evaluation.