A fuzzy spatial coherence-based approach to background/foreground separation for moving object detection

摘要: The detection of moving objects from stationary cameras is usually approached by background subtraction, i.e. constructing and maintaining an up-to-date model the detecting as those that deviate such a model. We adopt previously proposed approach to subtraction based on self-organization through artificial neural networks, has been shown well cope with several known issues for maintenance. Here, we propose spatial coherence variant enhance robustness against false detections formulate fuzzy deal decision problems typically arising when crisp settings are involved. show experimental results comparisons higher accuracy values can be reached color video sequences represent typical situations critical object detection.