The development of intelligent transport systems is aimed at providing real-time traffic information services to all road users, from autonomous vehicles, pedestrians, to the most vulnerable users, such as people with disabilities. In an ecosystem where vehicles will communicate with other vehicles and with different road infrastructures, it becomes crucial that vehicles are able to analyse and process the information received in order to make decisions. However, features such as high vehicle speeds, constant mobility, varying traffic density, etc., pose major challenges.  

The paper by Daniel Gehrig and Davide Scaramuzza presents an interesting advance in automotive vision using event cameras combined with RGB cameras that enhance real-time perception with high resolution and low latency, which is crucial for fast and accurate object detection in conditions of high speed and abrupt illumination changes. The authors use a convolutional neural network (CNN) to process images and an asynchronous graph neural network (GNN) to process events.