To extend and apply MVCC to more practical situations, in this paper we propose calibration-free multi-view crowd counting (CF-MVCC), which obtains the scene-level count directly from the density map predictions for each camera view without needing the camera calibrations in the test. ...
Particle filters can also be adopted to use crowd-sourced fingerprinting maps for the recalibration [21] and to fuse PDR and positioning estimation data in order to determine and re-estimate the divergence of particle trajectories. These re-estimated trajectories can be adopted to update the radio...
by linking the known time-stamped user positions with the time-stamped measured sensor devices205data, calibration data points are formed. As more users cumulatively enter the system, the accumulative amalgamation of crowd-sourced, multi-users' calibration points forms a distribution over the targeted...