Two-dimensional Maximum Local Variation

broadbeansromanceAI and Robotics

Nov 18, 2013 (3 years and 6 months ago)

56 views


Two
-
dimensional Maximum Local Variation
based on Image Euclidean Distance for Face

Recognition


Abstract



We have implemented an efficient system to recognize faces from
images with some near real
-
time variations.
The face recognition cycle starts with
the capturing of an individual’s face (normally either by still cameras, surveillance
cameras or
web

cameras
captured pictures). Afterwards the captured image is
manipulated using a combination of algorithms and neur
al network technology to
match face prints against others stored in a populated back
-
end database.

As
comparisons are made, the system assigns a value to the comparison using a
predetermined threshold, a match is declared.
This form of face recognition is
used
today to make accurate facial matches. Currently, this technology is being
implemented in parks, airports, arenas and other popular tourist locations.



In this paper,

we propose a linear approach, called two
-
dimensional

maximum
local variation (2DMLV
), for face recognition.
This system “Eigen faces” takes
eight points on the face and performs some algebraic manipulations to secure a
match. Neural Networks is used to increase reliability and offer a accurate match.
Another form of facial recognition is

Euclidean Distance

matching

which uses
graphic algorithms to trace points for object recognition.






Existing System




Face is a complex multidimensional visual model and developing a
computational model for face recognition is difficult.



Facial
recognition systems are computer
-
based security systems that are able
to automatically detect and identify human faces.



These systems depend on a recognition algorithm.
Principal Component
Analysis (PCA) is a statistical method under the broad title of fac
tor
analysis.



The Trace transform is a very rich representation of an image and in order to
use it directly for recognition, one has to produce a much simplified version
of it.




This will not yield accurate recognition system.



Less accurate



It does not
deal with biometric characteristics.










Proposed System




In this paper,

we propose a linear approach, called two
-
dimensional

maximum local variation (2DMLV), for face recognition.



Another form of facial recognition is
Euclidean Distance

matching

which
uses graphic algorithms to trace points for object recognition.



The transformed probe image is then matched with the gallery images which
display neutral expression
.



Neural Networks is used to increase reliabi
lity and offer a accurate match
are used

as inputs for positive and negative values variations.



New test image is taken for recognition (from test dataset) and its face
descriptor is calculated from the Eigen faces found before.
















SOFTWARE REQUIREMENTS



Platform




: JAVA

(JDK 1.5)



Front End : JAVA Swing



Back End : MySql



IDE : NETBEANS 6.9



Operating System





: Microsoft Windows 2000 or XP



HARDWARE
REQUIREMENTS




Processor




: Pentium IV Processor



RAM




: 512 MB



Monitor




: 14” VGA COLOR MONITOR





Keyboard




: 104 Keys



Capturin
g Device
: Web cam