Abstract
|
|
The Free Food Cam captures images of the graduate student lounge and is
mainly intended to signal the availability of leftover food. However, the
data it captures can be used to invade the privacy of the unsuspecting
visitors caught in its cross-hairs. Supervisors can find out what staff
members are taking extended coffee breaks; faculty advisors can find out
what graduate students are spending an inordinate amount of time in the
lounge. Below is a sample image from the webcam.
|
Introduction
|
|
The graduate student lounge is a relatively popular area with Computer
Science faculty, staff and graduate students. The population is relatively
closed which opens the door to "intimate stalkers". The Free Food Cam is a
publicly accessible webcam that offers a continuous view of the lounge.
Despite its narrow range of vision, it offers good quality photographs,
making it possible to recognize faces.
|
Methods
|
|
Data from the Free Food Cam was captured using a program every 10 minutes,
over a week. The images were stored locally, named after the day and the
time of the data capture. The resulting pictures were analyzed manually to
count the number of persons visible in each frame, and determine whether
food was present at the time.
|
Results
|
|
The number of persons visible in each frame was manually counted. The data
points showing the same person or the same group of persons in consecutive
frames were identified. The results were plotted versus time. Analysis of
the results revealed that the peaks of activity in the lounge correspond
to the 12PM-4PM timeframe and do not correspond to the availability of
free food.
Click here for an enlarged image
of the number of people appearing in the lounge in 30 minute increments.
|
Privacy
|
|
There were several instances where the Free Food Cam captured images of
the same person in several frames, and enabled us to clearly identify the
person: in one case, a staff member is seen taking an extended coffee
break lasting over 30 minutes. In another instance, the same graduate
student is seen reading the newspaper for over 30 minutes.
Below are two images showing the same person after some time lapse;
the face has blotted out to help de-identify the person.
|
References
|
|
Camera: https://zimbs.srv.cs.cmu.edu/coke/ffc.html
Project Paper (PDF)
|
