Detection of change in the number of humans in a monocular image sequence for pedestrian motion tracking

Detection and analysis of trajectories are effective for providing services to pedestrians in public spaces. We have proposed a method to detect moving objects from a monocular image sequence with a normal mixture model. However, the method cannot accommodate a changing number of moving objects. As described herein, to overcome this shortcoming, we proposed two algorithms for counting number of moving objects automatically. They perform successive counting with detection of an increase or decrease of moving objects. We did an experiment using an image sequence from an actual environment to verify the availability of the proposed method.