Accurate abandoned and removed object classification using hierarchical finite state machine

作者：

Highlights：

• We propose a novel and accurate ARO classification method.

• We propose a hierarchical FSM consisting of pixel-, region-, and event-layers.

• State transition is done by the pre-trained SVM using 7 different input features.

• The proposed ARO method shows higher classification and low false alarm.

• The proposed ARO method can be applied to many practical applications.

摘要

•We propose a novel and accurate ARO classification method.•We propose a hierarchical FSM consisting of pixel-, region-, and event-layers.•State transition is done by the pre-trained SVM using 7 different input features.•The proposed ARO method shows higher classification and low false alarm.•The proposed ARO method can be applied to many practical applications.

论文关键词：Support vector machine,Hierarchical finite state machine,Pixel classification,Region classification,Event classification

论文评审过程：Received 4 December 2014, Revised 17 June 2015, Accepted 11 September 2015, Available online 9 October 2015, Version of Record 25 October 2015.

论文官网地址：https://doi.org/10.1016/j.imavis.2015.09.004