Fuzzy logic to track objects from MPEG video sequences

L. Rodriguez-Benitez, J. Moreno-Garcia, J.J. Castro-Sanchez, L. Jimenez

This work presents a tracking algorithm that takes as input a MPEG compressed video. Our algorithm makes use of fuzzy logic during the video data extraction, the segmentation of regions with similar motion parameters and finally, the tracking of real objects. The presented method is tested on tracking of vehicles in traffic scenes obtaining good results in complex situations.

PDF full paper