Method of object movement tracking in video surveillance systems based on time series analysis

The purpose of this research is to identify possibilities of object movement tracking in video surveillance systems with a time series analysis. The proposed method allows to transform a set of input video frames into a signal, that depends on frame content and that can be decomposed into several components, which can be used for object's movement analysis.

The method based on a time series analysis of a signal based on singular spectrum analysis (SSA) method, which can decomposed an input signal into several subseries: trend, periodic component or noise. The most significant part of the decomposition for this method is a periodic component, which can provide information about visual changes of phases of human movement.

Current talk defines goals, problems and expected results of research.