Abstract
|
|
The goal of this project is to prove that it is possible to remotely
monitor an employee's performance automatically using publicly
available web cameras, and that for this reason the privacy
implications of these cameras must be considered.
|
Introduction
|
|
Upon finding a web site that contained publicly available web cameras
throughout one particular workplace in Florida, it became apparent that the
employees might not actually be aware of the cameras, and even if they
were they might not understand the full significance of its presence.
Normally, employees have some sense that their job performance may be
being monitored at any point because they can see their superiors in
the workplace. However, with a web camera, employees can be monitored
automatically and surreptitiously.
|
Methods
|
|
In order to detect the presence or absence of a receptionist
automatically, I have written a program that examines one specific
region of the image (the region behind the receptionist's desk). After
computing the average and standard deviation images of a large number
of receptionist-free images, we examine other pictures to see if they
contain large regions of pixels that are multiple standard deviations
above average. This is our indication that the receptionist is present
and doing his or her job.
|
Results
|
|
Just by using the simplistic methods used in this experiment, we were
able to achieve accuracy in automatic evaluation of around 70%. While
this would not be enough to base a firing on, it certainly suffices to
prove the concept. With more advanced computer vision techniques, we
would more than likely be able to increase the accuracy.
|
Privacy
|
|
This is an important privacy issue because it affects the rights of
the employees. Employees should have the right to know when they are
being monitored and when the security of their job is possibly at
stake. It is my belief that employees should be allowed to approve any
cameras that are placed in the work place as well as being informed of
the potential for abuse.
|
References
|
|
Camera: https://66.35.183.170/view/view.shtml
Company: https://www.nhgainesville.com/
Project Paper (PDF)
|