摘要: This paper addresses the problem of automatic detection and prediction abnormal human behaviours in public spaces. For this propose a novel classifier, called N-ary trees, is presented. The classifier processes time series attributes like object position, velocity, perimeter area, to infer type action performed. innovative can detect three types events: normal; unusual; or events. In order evaluate performance trees we carry out preliminary study with 180 synthetic tracks one restricted area. results revealed great level accuracy that proposed method be used surveillance systems.
参考文章(3)
A. Mecocci, M. Pannozzo, A. Fumarola, Automatic detection of anomalous behavioural events for advanced real-time video surveillance international conference on computational intelligence for measurement systems and applications. pp. 187- 192 ,(2003) , 10.1109/CIMSA.2003.1227225
Ron Kohavi, A study of cross-validation and bootstrap for accuracy estimation and model selection international joint conference on artificial intelligence. ,vol. 2, pp. 1137- 1143 ,(1995)
A. H. TICKNER, E. C. POULTON, Monitoring up to 16 Synthetic Television Pictures Showing a Great Deal of Movement. Ergonomics. ,vol. 16, pp. 381- 401 ,(1973) , 10.1080/00140137308924529