The Software utility constructed for this project has the three main functions:
Video Analysis – given a set of objects and a behavior model, the analyzer can recognize events that are taking place in a video.
It displays the recognized events in a log, which can then be saved to an xml document.
Video Annotation- given a set of objects and a video, allows the user to create a new annotation with new events. It is also possible to load an existing annotation from file. Existing events may be edited or deleted
Each event is defined by the attributes: name, start frame, end frame, the object IDs of objects participating in the event, and a free-text description.
At any point, it is possible to view the video with event and object annotations appearing on the video screen, as defined by the annotator interface. Saving the annotation to a file, in an xml format is possible. Both video analyzer and video annotator, generate the same xml format.
Annotation File comparison- given a ground truth event annotation file and an event annotation file from the annotator, this interface allows comparison between the two. The comparison is determined in a metric file, which defines the parameters of each event to compare by. The results of the comparison are displayed in a table on the screen.