|
计算机应用 2007
A novel trajectory pattern learning method based on vector quantization and depth first search
|
Abstract:
This paper puts forward a novel depth first search method to learning the distribution of motion trajectory based on the vector quantization of the flow vectors.Then the sequential patterns graph is generated,which visually represents the trajectory pattern.The corresponding anomaly detection method is also given in this paper.The experiments on different sites demonstrate that our method can not only discover the flow vector's distribution but also can reflect their time orders effectively, which makes it suitable for anomaly detection in outdoor scenes.