A Visual Recognition Security System in Java
Keiron Skillett, MEng Project 2001

 

Design Specification

2.1 Motivation

Millions of people all over the world log-on to websites daily, by entering a username and password, but how secure is this mechanism for entering websites?

Passwords are easily, forgotten, copied (more often than not, users will write down their password somewhere), guessed (name of girlfriend, birthday, street name etc). These are thing that the user has some control over, what about the hackers? Hacking into a database to copy a list of passwords? Changing a password in a database?

For example, on 1st December 2000, a 20-year-old man pleaded guilty to breaking into two computers owned by NASA's Jet Propulsion Laboratory in 1998 and using one to host Internet chat rooms devoted to hacking. NASA should surely be one of the most secure systems in the world, but a password was intercepted, leaked or found, allowing this young man to play inside a United States Government secure site.

What if, the only password into a website (or any computer system) was your face, or a simple object? To log on in the morning a user holds up their object to their web-cam (which come as standard on most machines now)? That object could be their calculator, mobile phone, a picture of their signature, a can of deodorant, a comb, even a strange shape drawn on a piece of paper, as long as that object was stored in the database the user would be authenticated.

It is obvious that more secure methods than this already exist:

  • Fingerprint and Retina scanners.

  • Swipe Cards

But these methods also have problems; both would require expensive hardware installation on every machine that required access to the system. Fingerprint and retina scanners are very secure, but swipe cards are easily lost, stolen or misplaced.
 

2.2 Objectives

Project 1 - 4th Year MEng Project Aims

This will be a Java application, which will listen for image data being sent to it over a network and then process it as necessary on the server before sending a response back to the client applet/application.

  • Upon receiving data, load into application.
  • Process data.
  • Image Analysis.
  • Send back response

Project 2 - 3rd Year BEng Project Aims

This should be either an applet or application on a client machine which has a camera attached.

  • Capture Image. (Prompt for object to be moved if in incorrect position).
  • Transmit data to recognition server.
  • Receive image data from client on server.
  • Possibly forward to administrator for review?
  • Transmit response from server to client
  • Wait for response from recognition server.
  • Display response from recognition server.

The BEng Student should decide any further details.

These two projects are completely exclusive to each other,
Project 1 can perform upon it's own as an image analysis/comparison application.
Project 2 can perform upon it's own as image capture software.

MEng Project Timetable

Figure 2.1 – MEng Project Timetable



Figure 2.2 - Division of work between the two projects.