Dan Pineo: Leveraging Deep Learning

by Dan Pineo on 02/10/2020

As the Director of Algorithm for the ProVision body scanner I am responsible for ensuring it detects potential threats consistently and quickly, with minimal false alarms. The scanner utilizes millimetre waves similar to those from a wireless router which pass through clothing and reflect off objects such as guns, knives and explosives. By tactically transmitting these waves from several precise positions and measuring the reflections we can determine the shape of any object.

Prior to joining L3Harris, I was developing advanced AI algorithms for prototype research systems used in government research labs. I was leveraging the latest AI algorithms and frameworks allowing me to gain valuable experience that I would later apply to the ProVision project. When I first started at L3Harris, I reviewed the algorithms that were being used and saw right away that they could benefit from using some of the recent advancements in the field of artificial intelligence, specifically a subfield called deep learning. Deep Learning  is a form of machine learning algorithms that process raw data and create patterns for use in decision making. I had been using it for years so I knew how to make the bespoke customizations needed to get the most out of the ProVision.

I knew this would constitute a major overhaul of the algorithm chain, but everyone at L3Harris was super supportive of the new direction. My team embraced deep learning wholeheartedly and quickly became leading AI experts within the company. It has to be said that my leadership team was very supportive and it was clear that they had full confidence in us in order to make such significant changes to the algorithms.

Our work modernizing the algorithms with deep learning really looks to be paying off. We’ve been seeing huge improvements in performance and customers are purchasing these software upgrades for their ProVision 2 systems. We recently created an algorithm that has passed ECAC’s upcoming 2.1 certification levels, something that has been out of reach for years, and we passed it with huge margins. The support from the people within the company has enabled us to breakthrough and it is great to be able to reward them for the confidence that they had invested in us.