Transcript fred_the_token

Face Recognition and
Biometric Filters
By Fred_the_token
Identity Confirmed:
Osama bin Laden
Basic Structure
Basic Functionality
• Image Input
– Still Image (JPEG, GIF)
• Facial Location
– Neural Network
• Facial Recognition
– Open Source From CSU
• Identity Output
Desired Options
• Video Input
– AVI, MPEG
• Platform Independence
– Everything but the GUI will be straight C++
code, and hopefully the GUI library will be
portable as well
Bells and Whistles
• Live Video Input
– Web Cam or Security Camera
• Optimized for Speed
• Ported to Windows 2000/XP
• Configuration Script for Easy Install
– With Help Files
Similar Systems
• Viisage
– Good: Very Fast
– Bad: Closed Source and Expensive
• OpenCV by Intel
– Can’t tell the web sight has been down for over a week
• Sluggish Software
– Good: Open Source
– Bad: Slow and Not Portable (Visual Basic)
• CSU Facial Recognition
– Good: Open Source for Academia
– Bad: Not Modular, Not Completely Open
What We Do Different
Ours is a System of Biometric Filters; It is
designed to be simple to use, customizable, Open
Source, and platform independent.
The filters will be designed so that they can be
multithreaded on a multi-processor system.
These filters can be implemented in hardware or
on a Digital Signal Processor (DSP). Thus they
can be optimized for speed and also embedded
into various systems.
Possible Customers
Academia
Security Professionals
Governments
User Interface Designers
The Super Bowl
Questions?