Figure 19
From: Automatic Reasoning about Causal Events in Surveillance Video
The text commentary for a person crossing the road at a set of traffic lights provides the input to the reasoning engine. Here, there is no interaction between people, but domain knowledge allows the system to recognise that the person walked onto the road, in order to cross the road. The same rules and events set as used to generate the results in Figure 18 is successfully used here in a different scene.