CAMeL: A Self-Adaptive Framework for Enriching Context-Aware Middlewares with Machine Learning Capabilities

Figure 3

An automaton driving the reconfiguration of a surveillance camera for automatic crowd detection. State A is used for detecting the presence of people or not. In state B people are counted, while state C represents the state in which the system has detected a crowd. Labels over the edges represents the keywords (i.e., the labels produced by active classifiers) triggering transitions.