Closed scorpeeon closed 10 years ago
Background model, processes the video sequence at pixel level and frame level (with Markov Random Field) alternatively. It can capture both the temporal and spatial context of the video sequence. Needs offline learning. Focuses only on problems on freeways.
ANovelProbabilityModelForBackgroundMaintenanceAndSubtraction.pdf