Camera motion detection is essential for automated video analysis. We propose a new probabilistic model for detecting zoom-in/zoom-out operations. The model uses EM to estimate the probability of a zoom versus a non-zoom operation from standard MPEG motion vectors. Traditional methods usually set an empirical threshold after deriving parameters proportional to zoom, pan, rotate and tilt. In contrast, our probabilistic model has a solid probabilistic foundation and a clear, simple probability threshold. Experiments show that this probabilistic model significantly out-performs a baseline parametric method for zoom detection in both precision and recall. © 2002 IEEE.
CITATION STYLE
Jin, R., Qi, Y., & Hauptmann, A. (2002). A probabilistic model for camera zoom detection. Proceedings - International Conference on Pattern Recognition, 16(3), 859–862. https://doi.org/10.1109/ICPR.2002.1048160
Mendeley helps you to discover research relevant for your work.