Dynamic segmentation using a novel neuromorphic polarimetric imaging system

The principles of a novel neuromorphic spike event-based polarimetric Dynamic Vision Sensor (DVS), mimicking the biological retina, for dynamic imaging of a scene aimed at increased temporal contrast, are introduced. A novel dynamic segmentation of moving objects was performed using a neuromorphic camera paired with a linear polarizingfilter. High-temporal contrast images were acquired and their local spatial-temporal content was further analyzed using wavelet decomposition. The outcome of the proposed technique clearly indicates that neuromorphic imaging coupled with polarimetric principles can identify and discriminate structures of different origin.