#162: Raul Bravo of Outsight

Dizzying progress in computer vision is one of the factors that has brought autonomous driving within reach, but as self-driving cars get closer to reality the probabilistic nature of machine learning inference is creating challenges due to the need for extremely high levels of safety. One possible approach to providing a more deterministic "safety checker" in future autonomous drive stacks comes from a company called Outsight, which is developing an active hyperspectral imaging sensor which it calls a "3D Semantic Camera" that holds the promise of (among other things) rapidly classifying objects based on spectral measurements of their material composition... without inference. Outsight co-founder Raul Bravo joins the show to explain how this sensor could be used to solve some of the trickiest perception-level challenges in autonomy, and potentially offer a powerful new tool to autonomous drive system developers.