Method, computer program product, and system for use with an environment divided into a plurality of predefined regions. The method comprises acquiring first image information including a first person, determining location information for the first person, and identifying a first behavior of the first person from a plurality of predefined behaviors. The method further comprises performing a first predefined action responsive to identifying the first behavior. Performing the first predefined action comprises determining a first region of the plurality of predefined regions corresponding to the location information, and accessing a memory storing predefined relationship information between the plurality of predefined behaviors and the plurality of predefined regions. The predefined relationship information comprises a plurality of predefined actions. Performing the first predefined action further comprises selecting, using at least the first behavior and the first region, the first predefined action from the plurality of predefined actions.