In this setting, people move around and interact in a public space and we want to identify
complex activities, such as people walking together, fighting or meeting. The aim is to
support security procedures in a public space, by identifying problematic events, such as
a fight.
Data
The data on which we build our identification is a set of recognized elementary (short-term)
behaviors, such as a person running, moving abruptly or a person standing still.
The data were provided from the benchmark
CAVIAR dataset.
Example
For our scenario we gather data about two persons (George and Alex) for specific moments in time, indicated
by numeric timestamps.
For each timestamp we hold the current person action status and the position in the space.
Below we describe the gathered data both in everyday language and in the corresponding formal
representation.
Plain Representation
Formal Representation
For the first timestamp (400) George's action in an abrupt motion
happensAt(abrupt(George), 400)
and we have located him in position (262, 285) of the camera picture co-ordinates.
holdsAt(coord(George)=(262, 285))
For the same timestamp (400) Alex's action in an abrupt motion
happensAt(abrupt(Alex), 400)
and we have located him in position (260, 288) of the camera picture co-ordinates.
holdsAt(coord(Alex)=(260, 288))
In the next timestamp (440) George keeps move abruptly
happensAt(abrupt(George), 440)
and we have located him in position (262, 286) of the camera picture co-ordinates.
holdsAt(coord(George)=(262, 286))
In the timestamp (440) Alex's action change in an active motion
happensAt(active(Alex), 440)
and we have located him in position (262, 285) of the camera picture co-ordinates.
holdsAt(coord(Alex)=(262, 285))
In the next timestamp (480) George's action change in an active motion
happensAt(active(George), 480)
and we have located him in position (262, 285) of the camera picture co-ordinates.
holdsAt(coord(George)=(262, 285))
In the same timestamp (480) Alex's action change again in an abrupt motion
happensAt(active(Alex), 480)
and we have located him in new position (267, 285) of the camera picture co-ordinates.
holdsAt(coord(Alex)=(267, 285))
Finally in timestamp (520) George's action change again in an active motion
happensAt(active(George), 520)
and we have located him in same position (262, 285) of the camera picture co-ordinates.
holdsAt(coord(George)=(262, 285))
In the timestamp (520) Alex's action stay in an active motion
happensAt(active(Alex), 520)
and we have located him in new position (262, 284) of the camera picture co-ordinates.
holdsAt(coord(Alex)=(262, 284))
2. Embed your knowledge
Description
In this step embed your knowledge rule in the system.
That rules reflecting on the different actions.
Example
For example we provide formal descriptions of what
constitutes that two persons fighting,
they are close or move abruptly.
The output of the process allows a quick glimpse
over a huge amount of data gathered within the day,
summarizing the main events and problems that help
you make decisions.
Now, you can provide this analysis as a service
to others, or use the findings within your
organization to do things, such as optimizing
procedures, improving customer experience or
minimizing risks.