You are on page 1of 5

Significance of the Study

For the past two decades, there has been several research concerning the ability of computers to
imitate human in recognizing faces. Moreover, numerous algorithms have been used over the
years for the advancement of face recognition. Face recognition are done from still images and
from captured images from video. Still images that have been used for face recognition are
photographs, mug shots, and facial composite sketches. Automatic face recognition has been
used in areas such as law enforcement, video surveillance, and security system access
authentication. This study will used facial composite sketch as the basis of the face recognition.
The study also focuses in the area of law enforcement and video surveillance.
The outcome of the study will be a great significant in law enforcement. The video surveillance
using CCTV camera is the important tool in detecting human faces that will be match to the faces
of the suspects(subjects) based from the composite sketches. The study can assist law
enforcement because it can help police in tracking the location of the suspects(subjects) sooner
than the usual.
The purpose of this study is to contribute to the advancement of video surveillance of CCTV
systems by integrating it with face sketch recognition software. Because CCTV nowadays are
just limited to the face recognition software based from the database of face photographs.
Moreover, face sketch recognition are just being done on the database of mug shots. The study
provide sufficient details about implementing and exhibiting the capabilities of support vector
machine incorporated in the face sketch recognition software.

Scope and Delimitations


The study involves facial composite sketch recognition software integrated with
CCTV system. In testing the study, the input to the software will not be composite
sketches of real suspects. The subject of the test will be composite sketches, still
done by a forensic artist, of persons ranging from the age of 16-50 years old. The
subject will be composed of males or females. The test will be done here in Far
Eastern University- Institute of Technology building. The software will be integrated
in the CCTV system of the said Universities. Therefore the specific type of CCTV
cameras to be used are the CCTV cameras used in the said Universities. Cameras
should be set in a manner that the frontal face of people walking will be shown. The
study will be limited to detecting faces that shows atleast 80% of human face.
Persons walking with their head bowed down or sideward will not be detected in this
study. The most important facial features such as the nose, the eyes, the mouth
and the ear should be visible. Frontal view are highly recommended. The study
is also limited scenes that has appropriate lightning that can show the facial
features of the person in the videos. For example, at daytime and/or at nighttime
that has sufficient light that clearly shows the facial feature of the human face
shown in the CCTV. The study will limit the detection of one subject at a time from a
database of composite sketch that contains 10 composite sketches. The frame of
the CCTV can contain atmost 20 persons to be able to detect a subject. Moreover,
the subject should be seen on the CCTV for atleast 5 seconds to be able to detect
and recognized it from the database of the composite sketch.

http://www.fedtechmagazine.com/article/2013/11/4-limitations-facial-recognitiontechnology

Conceptual Framework

Input

Static: Facial Composite Sketch


Dynamic: Captured Image of Detected face from CCTV

Process

Digitalizing Facial Composite Sketch Converting to JPEG Image


Image Enhancement of Facial Composite Sketch Image
Image Enhancement of Captured Image of Detected face from CCTV
Face Recognition of enhanced detected face image and facial composite sketch
through Support Vector Machines

Output
If the facial features of the detected face on the CCTV camera matched the facial
features of the facial composite sketch, system will alarm.
A message containing the time and date the images matched will be shown
If there is no match
Will not alarm
Will return to the process of Face Detection

IV. Conceptual Framework


Static Input:
Facial Composite Sketch

Sketch to Photo (JPEG)


Image Enhancement
Face Detection
Dynamic Input:
Captured Image of Detected face
from CCTV
Image Enhancement
Face Recognition:
Support Vector Machines
Is there a match between the facial features of the detected face seen on CCTV and
the facial features of a facial composite sketch stored in the database?
No
Yes

Alarm

Show log Message

Store the matched Images

Figure 1. Conceptual Framework


The system has two inputs; dynamic and static input. The dynamic input is the
captured image from the detected face on the CCTV camera, while the static input
is the facial composite sketch images in the database.
Facial composite sketches will be digitalized converting it to JPEG image. The facial
composites

sketch

images

will

undergo

preprocessing

stage

of

image

enhancement. The captured image of detected face from CCTV will also undergo a
preprocessing stage of image enhancement. Then, enhanced detected face image
and facial composite sketch will be matched through face recognition. The facial
features of the detected face on the CCTV camera will be matched to the facial
features of the facial composite sketch, through the algorithm of Support Vector
Machines.
If the facial features of the detected face seen on CCTV matches the facial features
of a facial composite sketch stored in the database, the system will output an alarm,
otherwise the system will not alarm. Also, a message containing the time and date
the system matched the images will be shown. Then the detected face that has a
matched in the facial composite sketch will be stored. After the system alarm or did
not alarm, the process of face detection and face recognition will be repeated.

You might also like