A Review on Feature Extraction Techniques and General Approach for Face Recognition
1 like199 views
The document reviews various feature extraction techniques for face recognition, highlighting their importance in combating identity fraud. It discusses methods such as PCA, LDA, and SURF, along with specific facial features like lips and eyes that can be extracted for improved recognition accuracy. The paper emphasizes the evolving nature of face recognition technology and its future potential as advancements in science and technology continue.
![International Journal of Computer Applications Technology and Research
Volume 5– Issue 3, 156 - 158, 2016, ISSN:- 2319–8656
www.ijcat.com 156
A Review on Feature Extraction Techniques and
General Approach for Face Recognition
Aakanksha Agrawal
Department of CSE
RCET, Bhilai
CG, India
Steve Samson
Department of CSE
RCET, Bhilai
CG, India
Abstract: In recent time, alongwith the advances and new inventions in science and technology, fraud people and identity thieves are
also becoming smarter by finding new ways to fool the authorization and authentication process. So, there is a strong need of efficient
face recognition process or computer systems capable of recognizing faces of authenticated persons. One way to make face recognition
efficient is by extracting features of faces. Several feature extraction techniques are available such as template based, appearance-
based, geometry based, color segmentation based, etc. This paper presents an overview of various feature extraction techniques
followed in different reasearches for face recognition in the field of digital image processing and gives an approach for using these
feature extraction techniques for efficient face recognition.
Keywords: face recognition, feature extraction, lip extraction, eye extraction
1. INTRODUCTION
Feature Extraction is a kind of process for reducing
dimensionality so as to represent the interesting image parts
with great efficiency as a compact feature vector [16]. This is
useful in case of large images.
No one can say which algorithm is best suitable for extracting
features. The algorithm selection is dependent on: (1) What
exactly is the task it needs to perform (2) Whether supervised
method is needed or unsupervised (3) Whether Inexpensive
method is required or strong computational method is required
etc. Some techniques for feature extraction are Speeded Up
Robust Features (SURF), Histogram of Oriented Gradients
(HOG), Color Histograms, Local Binary Patterns (LBP), Haar
wavelets, etc. [16]
Many software packages for data analysis are available
providing feature extraction and dimension reduction. Some
numerical programming environments such as MATLAB,
NumPy, SciLab also provides technique for feature extraction
like Principal Component analysis (PCA) etc. [18]
2. FACE FEATURES FOR
EXTRACTION
2.1 Lip Feature
Lip is a sensory organ existing in visible portion of human
face and considered to be different for each and every
individual. There are researches carried out for face
recognition and classification of gender using lip shape and
color analysis. [3]
2.2 Eye/Iris Feature
Eye/Iris has randomness due to its small tissues which
provides differentiation to the pattern of eye for each and
every individual human being. [13] The stableness,
uniqueness and non-invasion, these qualities make the iris
outstanding among several biometric features. [14]
2.3 Nose Feature
The nose tip is a distinctive point of human face. It also
remains unaffected even due to changes in facial expressions.
[18] Thus, it proves to be efficient for face recognition.
3. TECHNIQUES FOR FEATURE
EXTRACTION PURPOSE
3.1 Face Part Detection (FPD)
Convolution technique is used in this algorithm. It works by
multiplying vectors and returns values by using length and
width. Gabor features are done using Gabor filters and here
image decomposition is done by converting real part and
imaginary part. [12]
3.2 Principal Component Analysis (PCA)
In this, the dimensionality is reduced by projecting the data
onto the largest eigenvectors. [18] It selects the weights on the
basis of frequency in the frequency domain. It cannot separate
the class linearly.](https://image.slidesharecdn.com/ijcatr05031008-160420092701/85/A-Review-on-Feature-Extraction-Techniques-and-General-Approach-for-Face-Recognition-1-320.jpg)
![International Journal of Computer Applications Technology and Research
Volume 5– Issue 3, 156 - 158, 2016, ISSN:- 2319–8656
www.ijcat.com 157
3.3 Linear Discriminant Analysis (LDA)
LDA is a generalization of Fisher’s linear discriminant to find
a linear combination of features that characterizes two or more
classes of events. The resulting combination may be used for
dimensionality Reduction or as a linear classifier. [18]
3.4 Speeded Up Robust Features (SURF)
Surf algorithm is an improvement of Scale Invariant Feature
Transform (SIFT) algorithm. Surf uses a fast multi-scale
Hessian keypoint detector that can find keypoints. It can also
be used to compute user specified keypoints. Only 8 bit
grayscale images are supported. [3]
4. RELATED WORK
4.1 Rutuja G. Shelke, S.A.Annadate presented a novel
approach for Face Recognition and Gender classification
strategy using the features of lips. Here feature extraction is
carried out by using Principal component analysis (PCA) and
Gabor wavelet. Out of two techniques, results of Gabor filter
are more accurate and fast because it is having less leakage in
time frequency domain. [International Journal of Innovation
and Scientific Research (IJISR), Vol 10, No.2, Oct.2014,
Innovative Space of Scientific Research Journals (ISSR)].
4.2 Ishvari S. Patel, Apurva A. Desai used Preprocessing
techniques like Edge Detection by Canny Method and Height
and Width comparison for Lip Contour Detection. This model
works effectively and gives around 98% result for image
sequences but we can still improve accuracy of result by
extracting perfect lips region. [International Journal of
Scientific Research (IJSR), Volume II, Issue V, May 2013].
4.3 Sambit Bakshi, Rahul Raman, Pankaj K Sa paper
proposes that grayscale lip images constitute local features.
The claim has been experimentally established by
extracting local features applying two techniques viz. SIFT
and SURF. The results obtained are enough to establish that
unique local features exist in lip images through which an
individual can be recognized. [India Conference (INDICON),
2011 Annual IEEE].
4.4 Sasikumar Gurumurty, B. K. Tripathy divided
methodology into: Mouth Region Localization and Key
point‘s Extraction and Model Fitting. In first and second
steps, mouth region and key points are found by using hybrid
edges, which combine color and intensity information. In third
step, cubic polynomial models are fitted using key points
position and hybrid edges. An automatic, robust and accurate
lip segmentation method has been presented. This is
considered as good result and encourage for its use combined
with other biometrics systems.[ I.J. Intelligent Systems and
Applications (IJISA), July 2012 in MECS].
4.5 B. Sangeetha Devi, V.J.Arul Karthick used two processes
for lip recognition. First, face detection by Viola and Jones
algorithm. Second, lip detection by morphological operation
and five various mouth corner points. Lip biometric can be
used to authenticate an individual since the lip is unique.
[International Journal of Advanced Research Trends in
Engineering and Technology (IJARTET), Vol. II, Special
Issue I, March 2015].
5. GENERAL APPROACH FOR FACE
RECOGNITION
5.1 Acquiring the image of an individual’s
face
Digitally scan an existing photograph, or Acquire a live
picture of a subject.
5.2 Locate image of face
Software is used to locate the faces in the image that has been
obtained.
5.3 Analysis of facial image
Software measures face according to its peaks and valleys and
focuses on the inner area of the face.
5.4 Comparison
The face print created by the software is compared to all face
prints the system has stored in its database.
5.5 Match or No Match
Software decides whether or not any comparisons from the
above step are close enough to declare a possible match.
6. PERFORMANCE EVALUATION
PARAMETERS
6.1 False Acceptance Rate (FAR)
The probability that a system will incorrectly identify an
individual or will fail to reject an imposter.
(Also called Type2 Error Rate)
FAR=NFA/NIIA
NFA=number of false acceptance
NIIA=number of imposter identification attempts
6.2 False Rejection Rate (FRR)
The probability that a system will fail to identify an enrolee.
(Also called Type1 Error Rate)
FRR=NFR/NEIA
NFA=number of false rejection
NIIA=number of enrolee identification attempts
7. SCOPE OF FUTURE WORK
Face Recognition is a very vast and elaborated field. It has no
end. As the advancement in science and technology, new
techniques will continue developing day-by-day. Today, Lip
and Eye extraction are mostly discussed techniques for face
recognition purpose but in future many more advance
techniques will arise for performing Face Recognition with
much more accuracy and efficiency.
8. ACKNOWLEDGMENTS
Our sincere thanks to all the respected and experienced
faculties for their valuable guidance and motivation that
always encouraged us to give our full dedication towards a
new improvement in the field of science and image
processing.](https://image.slidesharecdn.com/ijcatr05031008-160420092701/85/A-Review-on-Feature-Extraction-Techniques-and-General-Approach-for-Face-Recognition-2-320.jpg)
![International Journal of Computer Applications Technology and Research
Volume 5– Issue 3, 156 - 158, 2016, ISSN:- 2319–8656
www.ijcat.com 158
9. REFERENCES
[1] Rutuja G. Shelke and S. A. Annadate, “Face Recognition
and Gender Classification Using Feature of Lips”,
International Journal of Innovation and Scientific Research
(IJISR), Innovative Space of Scientific Research Journals
(ISSR), Vol. 10, No. 2, Oct. 2014.
[2] B. Sangeetha Devi and V. J. Arul Karthick, “Lip
Recognition With Support Vector Machine (SVM)
Classifier”, International Journal of Advanced Research
Trends in Engineering and Technology (IJARTET), Vol. II,
Special Issue I, March 2015.
[3] Sambit Bakshi, Rahul Raman, Pankaj K Sa, “Lip pattern
recognition based on local feature extraction”, India
Conference (INDICON), IEEE Annual, 2011.
[4] Sunil Sangve, Nilakshi Mule, “Lip Recognition for
Authentication and Security”, IOSR Journal of Computer
Engineering (IOSR-JCE) Volume 16, Issue 3, Ver. VII, May-
Jun. 2014.
[5] Ishvari S. Patel and Apurva A. Desai, “Lip Segmentation
Based on Edge Detection Technique”, International Journal of
Scientific Research (IJSR), Volume II, Issue V, May 2013.
[6] Duy Nguyen, David Halupka, Parham Aarabi and Ali
Sheikholeslami, “Real-Time Face Detection and Lip Feature
Extraction Using Field-Programmable Gate Arrays”, IEEE
Transactions on Systems, Man and Cybernetics-Part B:
Cybernetics, Vol. 36, No.4, August 2006.
[7] Sasikumar Gurumurty and B. K. Tripathy, “Design and
Implementation of Face Recognition System in Matlab Using
the Features of Lips”, I.J. Intelligent Systems and
Applications (IJISA), July 2012 in MECS.
[8] Jyoti Bedre and Shubhangi Sapkal, “Comparative Study of
Face Recognition Techniques”, Emerging Trends in
Computer Science and Information Technology
(ETCSIT2012), International Journal of Computer
Applications (IJCA), 2012.
[9] Riddhi Patel and Shruti B. Yagnik, “A Literature Survey
on Face Recognition Techniques”, International Journal of
Computer Trends and Technology (IJCTT), Volume 5, No.4,
Nov 2013.
[10] John Daugman, “How Iris Recognition works”, IEEE
Transactions on Circuits and Systems for Video Technology,
Volume. 14, No. 1, Jan 2004.
[11] Mayank Vatsa, Richa Singh, and Afzel Noore,
“Improving Iris Recognition Performance Using
Segmentation, Quality Enhancement, Match Score Fusion,
and Indexing”, IEEE Transactions on Systems, Man and
Cybernetics— Part B: Cybernetics, Feb 2008.
[12] Dr. S. Vijayarani, S. Priyatharsini, “Facial Feature
Extraction Based On FPD and GLCM Algorithms”,
International Journal of Innovative Research in Computer and
Communication Engineering, Vol.3, Issue 3, March 2015.
[13] Pankaj K.Sa, S.S. Barpanda, Bansidhar Manjhi, “Region
Based Feature Extraction from Non-Cooperative Iris Images”,
Innovation Syst Software Engg, 2015.
[14] Changcheng Li, Weidong Zhou, Shasha Yuan, “Iris
Recognition Based on a Novel Variation of Local Binary
Pattern”, Springer-Verlag Berlin Hiedelberg, Vis Comput,
2014.
[15] Rafael C. Gonzalez, Richard E. Woods and Steven L.
Eddins, “Digital Image Processing Using MATLAB”, Second
Edition
[16] The MathWorks, Inc., “Image Processing Toolbox”,
User's Guide, COPYRIGHT 1993–2015
[17] S. N. Sivanandam, S. Sumathi, S. N. Deepa “Neural
Network using MATLAB 6.0”.
[18] www.google.com](https://image.slidesharecdn.com/ijcatr05031008-160420092701/85/A-Review-on-Feature-Extraction-Techniques-and-General-Approach-for-Face-Recognition-3-320.jpg)
A Review on Feature Extraction Techniques and General Approach for Face Recognition
- 1. International Journal of Computer Applications Technology and Research Volume 5– Issue 3, 156 - 158, 2016, ISSN:- 2319–8656 www.ijcat.com 156 A Review on Feature Extraction Techniques and General Approach for Face Recognition Aakanksha Agrawal Department of CSE RCET, Bhilai CG, India Steve Samson Department of CSE RCET, Bhilai CG, India Abstract: In recent time, alongwith the advances and new inventions in science and technology, fraud people and identity thieves are also becoming smarter by finding new ways to fool the authorization and authentication process. So, there is a strong need of efficient face recognition process or computer systems capable of recognizing faces of authenticated persons. One way to make face recognition efficient is by extracting features of faces. Several feature extraction techniques are available such as template based, appearance- based, geometry based, color segmentation based, etc. This paper presents an overview of various feature extraction techniques followed in different reasearches for face recognition in the field of digital image processing and gives an approach for using these feature extraction techniques for efficient face recognition. Keywords: face recognition, feature extraction, lip extraction, eye extraction 1. INTRODUCTION Feature Extraction is a kind of process for reducing dimensionality so as to represent the interesting image parts with great efficiency as a compact feature vector [16]. This is useful in case of large images. No one can say which algorithm is best suitable for extracting features. The algorithm selection is dependent on: (1) What exactly is the task it needs to perform (2) Whether supervised method is needed or unsupervised (3) Whether Inexpensive method is required or strong computational method is required etc. Some techniques for feature extraction are Speeded Up Robust Features (SURF), Histogram of Oriented Gradients (HOG), Color Histograms, Local Binary Patterns (LBP), Haar wavelets, etc. [16] Many software packages for data analysis are available providing feature extraction and dimension reduction. Some numerical programming environments such as MATLAB, NumPy, SciLab also provides technique for feature extraction like Principal Component analysis (PCA) etc. [18] 2. FACE FEATURES FOR EXTRACTION 2.1 Lip Feature Lip is a sensory organ existing in visible portion of human face and considered to be different for each and every individual. There are researches carried out for face recognition and classification of gender using lip shape and color analysis. [3] 2.2 Eye/Iris Feature Eye/Iris has randomness due to its small tissues which provides differentiation to the pattern of eye for each and every individual human being. [13] The stableness, uniqueness and non-invasion, these qualities make the iris outstanding among several biometric features. [14] 2.3 Nose Feature The nose tip is a distinctive point of human face. It also remains unaffected even due to changes in facial expressions. [18] Thus, it proves to be efficient for face recognition. 3. TECHNIQUES FOR FEATURE EXTRACTION PURPOSE 3.1 Face Part Detection (FPD) Convolution technique is used in this algorithm. It works by multiplying vectors and returns values by using length and width. Gabor features are done using Gabor filters and here image decomposition is done by converting real part and imaginary part. [12] 3.2 Principal Component Analysis (PCA) In this, the dimensionality is reduced by projecting the data onto the largest eigenvectors. [18] It selects the weights on the basis of frequency in the frequency domain. It cannot separate the class linearly.
- 2. International Journal of Computer Applications Technology and Research Volume 5– Issue 3, 156 - 158, 2016, ISSN:- 2319–8656 www.ijcat.com 157 3.3 Linear Discriminant Analysis (LDA) LDA is a generalization of Fisher’s linear discriminant to find a linear combination of features that characterizes two or more classes of events. The resulting combination may be used for dimensionality Reduction or as a linear classifier. [18] 3.4 Speeded Up Robust Features (SURF) Surf algorithm is an improvement of Scale Invariant Feature Transform (SIFT) algorithm. Surf uses a fast multi-scale Hessian keypoint detector that can find keypoints. It can also be used to compute user specified keypoints. Only 8 bit grayscale images are supported. [3] 4. RELATED WORK 4.1 Rutuja G. Shelke, S.A.Annadate presented a novel approach for Face Recognition and Gender classification strategy using the features of lips. Here feature extraction is carried out by using Principal component analysis (PCA) and Gabor wavelet. Out of two techniques, results of Gabor filter are more accurate and fast because it is having less leakage in time frequency domain. [International Journal of Innovation and Scientific Research (IJISR), Vol 10, No.2, Oct.2014, Innovative Space of Scientific Research Journals (ISSR)]. 4.2 Ishvari S. Patel, Apurva A. Desai used Preprocessing techniques like Edge Detection by Canny Method and Height and Width comparison for Lip Contour Detection. This model works effectively and gives around 98% result for image sequences but we can still improve accuracy of result by extracting perfect lips region. [International Journal of Scientific Research (IJSR), Volume II, Issue V, May 2013]. 4.3 Sambit Bakshi, Rahul Raman, Pankaj K Sa paper proposes that grayscale lip images constitute local features. The claim has been experimentally established by extracting local features applying two techniques viz. SIFT and SURF. The results obtained are enough to establish that unique local features exist in lip images through which an individual can be recognized. [India Conference (INDICON), 2011 Annual IEEE]. 4.4 Sasikumar Gurumurty, B. K. Tripathy divided methodology into: Mouth Region Localization and Key point‘s Extraction and Model Fitting. In first and second steps, mouth region and key points are found by using hybrid edges, which combine color and intensity information. In third step, cubic polynomial models are fitted using key points position and hybrid edges. An automatic, robust and accurate lip segmentation method has been presented. This is considered as good result and encourage for its use combined with other biometrics systems.[ I.J. Intelligent Systems and Applications (IJISA), July 2012 in MECS]. 4.5 B. Sangeetha Devi, V.J.Arul Karthick used two processes for lip recognition. First, face detection by Viola and Jones algorithm. Second, lip detection by morphological operation and five various mouth corner points. Lip biometric can be used to authenticate an individual since the lip is unique. [International Journal of Advanced Research Trends in Engineering and Technology (IJARTET), Vol. II, Special Issue I, March 2015]. 5. GENERAL APPROACH FOR FACE RECOGNITION 5.1 Acquiring the image of an individual’s face Digitally scan an existing photograph, or Acquire a live picture of a subject. 5.2 Locate image of face Software is used to locate the faces in the image that has been obtained. 5.3 Analysis of facial image Software measures face according to its peaks and valleys and focuses on the inner area of the face. 5.4 Comparison The face print created by the software is compared to all face prints the system has stored in its database. 5.5 Match or No Match Software decides whether or not any comparisons from the above step are close enough to declare a possible match. 6. PERFORMANCE EVALUATION PARAMETERS 6.1 False Acceptance Rate (FAR) The probability that a system will incorrectly identify an individual or will fail to reject an imposter. (Also called Type2 Error Rate) FAR=NFA/NIIA NFA=number of false acceptance NIIA=number of imposter identification attempts 6.2 False Rejection Rate (FRR) The probability that a system will fail to identify an enrolee. (Also called Type1 Error Rate) FRR=NFR/NEIA NFA=number of false rejection NIIA=number of enrolee identification attempts 7. SCOPE OF FUTURE WORK Face Recognition is a very vast and elaborated field. It has no end. As the advancement in science and technology, new techniques will continue developing day-by-day. Today, Lip and Eye extraction are mostly discussed techniques for face recognition purpose but in future many more advance techniques will arise for performing Face Recognition with much more accuracy and efficiency. 8. ACKNOWLEDGMENTS Our sincere thanks to all the respected and experienced faculties for their valuable guidance and motivation that always encouraged us to give our full dedication towards a new improvement in the field of science and image processing.
- 3. International Journal of Computer Applications Technology and Research Volume 5– Issue 3, 156 - 158, 2016, ISSN:- 2319–8656 www.ijcat.com 158 9. REFERENCES [1] Rutuja G. Shelke and S. A. Annadate, “Face Recognition and Gender Classification Using Feature of Lips”, International Journal of Innovation and Scientific Research (IJISR), Innovative Space of Scientific Research Journals (ISSR), Vol. 10, No. 2, Oct. 2014. [2] B. Sangeetha Devi and V. J. Arul Karthick, “Lip Recognition With Support Vector Machine (SVM) Classifier”, International Journal of Advanced Research Trends in Engineering and Technology (IJARTET), Vol. II, Special Issue I, March 2015. [3] Sambit Bakshi, Rahul Raman, Pankaj K Sa, “Lip pattern recognition based on local feature extraction”, India Conference (INDICON), IEEE Annual, 2011. [4] Sunil Sangve, Nilakshi Mule, “Lip Recognition for Authentication and Security”, IOSR Journal of Computer Engineering (IOSR-JCE) Volume 16, Issue 3, Ver. VII, May- Jun. 2014. [5] Ishvari S. Patel and Apurva A. Desai, “Lip Segmentation Based on Edge Detection Technique”, International Journal of Scientific Research (IJSR), Volume II, Issue V, May 2013. [6] Duy Nguyen, David Halupka, Parham Aarabi and Ali Sheikholeslami, “Real-Time Face Detection and Lip Feature Extraction Using Field-Programmable Gate Arrays”, IEEE Transactions on Systems, Man and Cybernetics-Part B: Cybernetics, Vol. 36, No.4, August 2006. [7] Sasikumar Gurumurty and B. K. Tripathy, “Design and Implementation of Face Recognition System in Matlab Using the Features of Lips”, I.J. Intelligent Systems and Applications (IJISA), July 2012 in MECS. [8] Jyoti Bedre and Shubhangi Sapkal, “Comparative Study of Face Recognition Techniques”, Emerging Trends in Computer Science and Information Technology (ETCSIT2012), International Journal of Computer Applications (IJCA), 2012. [9] Riddhi Patel and Shruti B. Yagnik, “A Literature Survey on Face Recognition Techniques”, International Journal of Computer Trends and Technology (IJCTT), Volume 5, No.4, Nov 2013. [10] John Daugman, “How Iris Recognition works”, IEEE Transactions on Circuits and Systems for Video Technology, Volume. 14, No. 1, Jan 2004. [11] Mayank Vatsa, Richa Singh, and Afzel Noore, “Improving Iris Recognition Performance Using Segmentation, Quality Enhancement, Match Score Fusion, and Indexing”, IEEE Transactions on Systems, Man and Cybernetics— Part B: Cybernetics, Feb 2008. [12] Dr. S. Vijayarani, S. Priyatharsini, “Facial Feature Extraction Based On FPD and GLCM Algorithms”, International Journal of Innovative Research in Computer and Communication Engineering, Vol.3, Issue 3, March 2015. [13] Pankaj K.Sa, S.S. Barpanda, Bansidhar Manjhi, “Region Based Feature Extraction from Non-Cooperative Iris Images”, Innovation Syst Software Engg, 2015. [14] Changcheng Li, Weidong Zhou, Shasha Yuan, “Iris Recognition Based on a Novel Variation of Local Binary Pattern”, Springer-Verlag Berlin Hiedelberg, Vis Comput, 2014. [15] Rafael C. Gonzalez, Richard E. Woods and Steven L. Eddins, “Digital Image Processing Using MATLAB”, Second Edition [16] The MathWorks, Inc., “Image Processing Toolbox”, User's Guide, COPYRIGHT 1993–2015 [17] S. N. Sivanandam, S. Sumathi, S. N. Deepa “Neural Network using MATLAB 6.0”. [18] www.google.com