Face Recognition

(Name of Author)

(Class/Subject)

(Teacher name)

March 4, 2019
Introduction

In this technological era, face recognition has been gaining much attraction and attention around the globe.
This technology benefits us in many different walks of life such as network security, data security,
avoidance of theft, retrieving data and much more. Not only this but hackers also have to face a great deal
of troubles because of face recognition technological advancements therefore, accessing control via face
recognition over any network increases the probability of user-friendliness and safety. Hence, passwords
are ‘Safe’ nowadays. In this research paper, focus has been made on this innovative technology, its
methods have been reviewed in detail and discussion has ben made on the recent advances in face
recognition technology.

It can be rightly observed that since we are living in a networked society; therefore, the need to maintain
the security of both physical or informational assets are in high demand which is somehow difficult too.
At different times, news is spread about the credit card theft, its fraud, scam, hacker’s hacking valuable
and confidential data and information, and/or breach of code of security in an organization or government
institution. A study shows that in the year 1998, some hackers caused $100 million US loss (Reuters,
1999). The point to ponder on is that these criminals get the benefits of loopholes in the access to control
systems because if we observe considerably we find that systems can be accessed through the assets we
carry such as ID cards, keys, password PINs, or some other relevant information. These can easily be
hacked and hackers can authenticate us or may be replicate us in the form of any confidential information
to be misused. Therefore, the need to truly verify a person is thought to be turned into reality which comes
now as “Biometrics” which helps identify and verify a person in terms of physiological features such as
fingerprints, face recognition, thumb impression, etc.1 Among these features some are considered as
highly reliable. For example, voice key or writing key stroke are less reliable because voice can be altered
due to any possible uncertain scenario such as sore throat; similarly, writing key strokes may be changed
because of some anxiety, over-excitement or stress whether positive or negative. On the other hand face
recognition will remain same unless some major injury happen to the person due to which the facial
features can be altered or changed or unidentified y the system. Hence, it has the accuracy without being
disruptive, that’s why since 1970s, it has captured researchers attention (Kelly, 1970).

1Kevin, Bowyer, Face Recognition Technology: Security vs Privacy, IEEE Technology and Society
Magazine, 2004)
How Does A “Face Recognition” System Work?
The basic operation procedure of a face recognition system is quite simple. First of all, a camera is installed
that can view some space such as the boarding area at the airport. The purpose is not to keep on having
the eyes stick on the computer monitor but rather let the monitor detects an interesting face and have an
alert system. It means the first step is when a computer detects a person’s face. Once captured, the image
of the face is cropped to checked for certain possible matches in the gallery.

Figure1: Front faces are detected only 2

Changes in the facial features in terms of expressions and hairstyles are yet another important concerns to
be taken care of for face recognition technology. There is a 2D mesh-like neutral network to overcome
this concern which possibly corrects changes in the facial expressions when recognizing the face (Lades,
1993). Another way is that instead of detecting the whole face of a person to be recognized later when
required, it is better to be as close to the specific face region as possible such as both eyes and the nose
except the mouth and ears. Figure 2 highlights the specific regions.3

2 Shang-Hung, Lin, An Introduction to Face Recognition Technology, (Part 2, Volume3 No 1, 2000)

3Lin, p.4, (Part 2, Volume3 No 1, 2000)
Recent Techniques in Face Recognition4
A. Line Edge Map (LEM)

This approach focuses on the face lines and can be taken as a combination
of two templates namely matching and geometrical face features. It provides
high accuracy recognition benefits with less memory requirements.

B. Support Vector Machine (SVM)

It is a techniques used for learning and is considered as an effective one to recognize face without requiring
and depending upon other data and knowledge. It is also important for minimizing risks of data theft and
hacking of the access of control.

C. Multiple Classifier Systems (MCSs)

It uses a combination of set patterns developing high accuracy and performance in the field of face
recognition.
A research Finding has been shown below to highlight the results

TABLE VIII
 COMPARISON OF DIFFERENT FACE DATABASES5

Database Description Limitation

AT&T Faces of 40 persons shot at Less persons, light and shades

(ORL PREVIOUSLY) different times with different issue
expressions

Oulu Physics Comprises of front facial No lighting variations

features leaving behind the side

4 Tolba, Ahmad, El-Baz, Ali, Face Recognition Review, 2005

5 Tolba, Ahmad, El-Baz, Ali, Face Recognition Review, 2005 p. 13
 line of a face

XM2VTS It uses sequences of 1000 Bytes Doesn’t give much information

videos and audio recordings about lighting and angle

Yale Grayscale images Limited persons

Review of Face Recognition Methods

This process of recognizing face is a three step process that involves face normalization, extracting
features, and comparison. It can be further categorized as based on the appearance or algorithms based on
certain designed models.

It is a profound fact that for safety reasons one cannot be asked to come and put his/her thumb impression
all the time, yet, surveillance can be made easy and accessible for later purpose in the form of face
recognition which is globally implemented for the purpose of keeping a record of people as well as to
retrieve the information and/or data required later on. Therefore, three different approaches are
implemented: appearance based, feature based and soft-computing based.

Appearance Based Recognition

It comprises of two important approaches namely Holistic approach and Hybrid approach; former uses
the whole face feature to be recognized while the latter can be extended to Eigen features such as eyes,
mouth, etc.
Fig. 2. Face represented by a small number of features.6

6
Roomi, Mansoor, Beham, Dr. M. Parisa, International Journal of Pattern Recognition and Artificial Intelligence,
2013.
Feature-based Face Recognition
It uses local features of faces to identify individuals. It comprises of eyes, nose, mouth, chin, and head
selected from different images. Certain topographical graphs are used in this approach to present link
between face-features and a graph scheme to keep a record of recognized person.

Soft Computing Face Recognition

It has emerged as an important way to recognize face for analyzing computer research in terms of
artificial intelligence. It has the disadvantage of uncertainty involved here.

Industry and Government Policies for Face Recognition Technology7

It is yet to be decided that what policy should be implemented for the technology of face recognition at
different public and government places. However, it is a profound fact that it provides security, yet,
another concept says that it intrudes privacy. Sometimes, criminal cases can be solved based on the face
recognition technology that helps the security institutions to resolve the matter at hand and to avoid any
unwanted circumstances. However, many studies focus on the point that face recognition technology can
also come up with the idea that a person’s data can be retrieved for later reference purposes for his own
safety; yet, on the contrary, there is a strong view as shown in the study conducted that quite a handsome
percentage of people consider it as an intrusion in their privacy if in case their pictures or images be used
unwisely or mishandled by others.

Some recent advances have also been recorded at certain places which includes eye retina scan, tongue
recognition scan but people are not that much happy with it. High probability can be found in their
accuracy and better performance in terms of safety, security and being well-informed. However, it has yet
to be decided that should all public places be installed with this latest technology to better avoid any bad
scenario and to keeping people safe after maintaining their face features record in the data base which can
be taken into consideration after a while also.

7 Bowyer, Kevin, Face Recognition Technology: Security vs Privacy, p.8)

Conclusion

A challenging and vital technique in today’s world is Face Recognition Technique which can be
considered inevitable in the world today because of uncertain life scenarios and unwanted situations all
are confronting with. However, it is also correct and be an undeniable fact that some people finds it
disturbing for them since it creates mental trauma for the m that the images taken at different places may
be misused or mishandled by others. Therefore, one cannot deny this fact that it also saves a person from
being threatened and black mailed. It also comprises of different approaches as discussed in the paper
above which can clarify that there are certain models and techniques to keep the information and data safe
and free from being hacked. Because of this technology hackers are really hanging in the balance for
thinking what to do with this technology since it has high probability of higher accuracy and better
performance of saving the data in most secured and un hacked manner. It is rather hoped that his paper
will provide important facts and details required to understand this technology and appreciate the readers
to find this topic interesting.

Therefore, it can be well-analyzed that this has been in high consideration for past few centuries and many
institutions are constantly working in this field to make it more reliable and note-worthy so that these can
be installed at public places. It is yet required that technology may make people realize the fact that they
are not only being watched rather they are safe.
 References

1. Bowyer, Kevin, Face Recognition Technology: Security vs Privacy, p.8)

2. D. Aradhana, K. Karibasappa and A. Chennakeshva Reddy, Face recognition using soft
computing tools: A survey, UbiCC J. 6(3) (2009) 854 863. 


3. Kevin, Bowyer, Face Recognition Technology: Security vs Privacy, IEEE Technology and
Society Magazine, 2004)
4. Lin, p.4, (Part 2, Volume3 No 1, 2000)

5. M. Agarwal, N. Jain, M. Kumar and H. Agrawal, Face recognition using eigen faces and arti ̄cial
neural network, Int. J. Comput. Theor. Eng. 2(4) (2010) 624 629. 


6. Roomi, Mansoor, Beham, Dr. M. Parisa, International Journal of Pattern Recognition and
Artificial Intelligence, 2013.

7. Shang-Hung, Lin, An Introduction to Face Recognition Technology, (Part 2, Volume3 No 1,

2000)
8. Tolba, Ahmad, El-Baz, Ali, Face Recognition Review, 2005
9. Tolba, Ahmad, El-Baz, Ali, Face Recognition Review, 2005 p. 13