Thermal image of a typical video-surveillance scene.
Projects
In this projects we investigated the complementary information given by different imaging modalities for the task of pedestrian/object detection.
In particular we combined optical, thermal, depth and motion cues to increase the detection and classification accuracy.
