Detecting Motion Regions Using Statistic Parameters

摘要

This chapter has shown a novel statistic parametric model for motion detection, incorporating the parameters of frequency, time center, and last time. In the model, the statistical parameters for a pixel sample sequence can be denoted in a quintuple style. SPM can be updated in time and accurately by doing statistical analysis for history samples. The experiments show that our presented model can suppress false detections from tail phenomenon, shadows, illumination small and abrupt change, repetitive motion, cluttered areas, and so on.