Human Detection, tracking and prediction using Laser scan + vision

edit

Hello guys. I am currently looking for approaches which uses fusion between laser scan and vision (RGB Camera) in order to detect, track and predict human motions, implemented on ROS. If anyone can share some robust approaches implemented in ROS? I am mainly interested to fuse the two types of data for better human motion prediction. For now on ROS wiki, i could only find cob_people_detection but this is not exactly fusing the the two sensors data rather it uses each of them separately. THanks