Table 5 Summary of existing works on using CEP for physical security purposes

[62]

Itria et al.

2014

Armed robbery, act of vandalism, dangerous object placed by person, melee.

Method for detecting critical situations that employs event correlation technologies to perform online analysis of real-world event via CEP engine such as Esper is proposed.

+ Event correlation technology is used to perform online analysis of real-world events via Esper.

- Rules are manually defined for each case.

[45]

Shahad et al

2018

Anomalous activities in smart building

CEP engine based on CAISER and NEsper are used to monitor and analyse complex event produced by CCTV and door sensors for activity recognition in smart building applications.

+ Comparison of two CEP engines, and different CE detectors using metrics of confusion matrix.

- Only person movement activity is considered, and no standard/public dataset was used.

[61]

Shahad et al.

2018

Break-ins, burglaries, and robberies

Suspicious loitering detection from annotated CCTV is proposed based on CEP approach (CAISER)

+ Performance of different CE detectors was evaluated with noisy and no-noise event annotation.

- Not implemented for real-time smart surveillance system.

[70]

Amru et al.

2018

Loitering ambient, falling and injured

Real-time event detection based on CEP (CAISER) for intelligent building surveillance system is proposed.

+ System was evaluated at different alignment situations in term of detection accuracy and ROC curve.

- Limited to single person tracking.

[71]

Mijović et al.

2019

Fire or terrorist attack

Intelligent and event driven layer powered by CEP is proposed to handle a flood of information coming from many sensors in critical infrastructures such as airports.

+ System allows operator to react automatically or manually to risk.

- No detail analysis of system performance in term of accuracy or comparison with different systems.