Current Internet of Things systems often incorporate the usage of advanced AI algorithms and DISTINGO is no exception. We gather great quantities of data from our locker cabinets and cells (usage patterns, occupation, ambient physical measurements) and the images from inside the cabinet cells and the surroundings of the cabinet are no exception. We gather images and video from both sides (all in accordance to the GDPR constraints in the EU) and feed them to different algorithms for processing. The exact nature of these algorithms will evolve, but basically we want to know what is happening inside (e.g., contents of the cell, whether a dangerous item such as a weapon or bomb has been left there) and outside of the full cabinet (e.g., what’s the headcount near the cabinet, how many people approach it, detection of vandalism to teh cabinet or the surroundings). For this, we use a combination of off-the-shelf algorithms (e.g., YOLO for object recognition and openpose to detect the persons around) and extend the networks with novel approaches to detect particular activities or objects. For this we’re extensively re-training the networks to suit better DISTINGO’s needs. This will be showcased in future demonstrators and videos, so stay tuned for more information.