Face recognition Face Identification

PROPOSED WORK ON
FACE RECOGNITION
Presented By
Kalyan Acharjya
A Presentation on Initial stage of M.Tech Dissertation Work
Department of…………………………
University of………..

2
There are some images (slide No 8) used within this presentation were
copied from internet without prior permission from original owner.
Only Original Owner has full rights reserved for copied images.
This PPT is only for fair academic use (Not Commercial).
Kalyan Acharjya
kalyan.acharjya@gmail.com
www.factsaboutuniversity.com
Disclosure

CONTENTS
 Introduction to Digital Image Processing.
 Face Recognition.
 Why Face Recognition.
 How Face Recognition.
 Literature Survey [Going On].
 Problem Statement.
 Challenges for Real Time Applications.
 Standard Face Images Database.
 Conclusions and Future Work.
 References.
3
Kalyan Acharjya, India

INTRODUCTION : DIGITAL IMAGE PROCESSING
 An Image has infinite intensity value.
 Also infinite picture point -How its stored?
 Digitization of image.
 Spatial discretization by Sampling.
 Intensity discretization by Quantization.
 An stored image is process in various means
(Techniques) for enhance or extracts some
features from it, is well considerable as
DIGITAL IMAGE PROCESSING. 4

FACE RECOGNITION
 How computer or systems is identify any person by comparison its
FACE with its previous stored database. Its also a part of
COMPUTER VISION.
 FACE RECOGNITION is the part of Digital Image Measurement.
 Its High Level Processing involved making sense of an ensemble of
recognize FACE with analysis of unknown FACE.
 Its almost similar to Human being, who identify any person, if
he/she have already met. Although it(Human) fails sometimes (rare
case).
5

WHY FACE RECOGNITION ?
 The world is urged for more and accurate face recognition rate.
 How COMPUTER VISITON is possible, as human being are?
 Automatic person identification.
 FACE RECOGNISITION have lots of real world applications.
 Automatic Attendance System.
 Security Purposes.
 Computer Interaction etc.
 Crowd Surveillance. [US (MIT) invested $ 100 million for perfect recognition
system-Times of India, Oct 2013 ].
 In 2011, London riots many suspects of partial face images were not
recognized by COTS FR system[15].
6

Comparison
HOW FACE RECOGNITION?
7
Input Image
Face Detection*
Crop Face Image
Features Extraction
Identification
Face Image
Database
*The targeted work will not include Face Detection Part.
*The input images will crop face images from standard Face Database.
Who is She?

LITERATURE SURVEY [1]
 Title : Li, Liao and Jain, “Partial Face Recognition,
Alignment free Approach”, IEEE, May, 2013.
 Technique Used: Authors proposed an alignment free
face recognition method based on multi-key point
descriptors. (MKD).
 Conclusion: Authors concluded that MKD method is
superior than leading commercial FR systems like
Pitpatt and faceVACS SDK.
8

9
LITERATURE SURVEY CONTD..[2]
 Title : Mersico, Nappi and Wechsler, “Robust Face
Recognition fro Uncontrolled Pose and Illumination ”,
IEEE, January, 2013.
 Technique Used: Authors proposed a novel frame work
based on normalization strategies and Face Analysis
for Commercial Entities (FACE).

 Conclusion: The result showed the significant increase
in recognition rate [95% in FERET fa Database] in
accuracy, whether comparison with other available
algorithms.

10
LITERATURE SURVEY CONTD.. [3]
 Title : Park and Savvides , “Individual Kernal Tensor
Subspaces for Robust Face Recognition: A Computationally
Efficient Tensor Framework without requiring Mode
Factorization ”, IEEE, Oct, 2007.
 Technique Used: The work based on high order tensor to
construct a multi linear structure and model the multiple
factors of face variations.
 Conclusion: The paper introduced the new concept that
appearance factor, the factor of person’s identity modeled
by a tensor structure can be used for better face
recognition system specially for difference types of
appearance of same faces.

11
LITERATURE SURVEY CONTD.. [4]
 Title: Karim, Lipu, Rahman and Sultana , “Face Recognition
using PCA based Method”, IEEE, 2010.
 Technique Used: The work based on Principle Component
Analysis (PCA).
 Conclusion: The paper concluded the Principle Component
Analysis is better then their predecessor, where recognition
rate 84.1 % (Male Face) and 95.45 % (Female Face) in case
of Indian face database.
 Also recognition rate 92.5 % (Male Face) and 85 % (Female
Face) in case of University of Essex, UK face database.

12
LITERATURE SURVEY CONTD..[5] MANY MORE…..!
 Title : Meade, Kumar and Phillips, “Comparative
performance of Principle Component Analysis , Gabor
Wavelets and Discrete Gabor Wavelets”, Canadian
Journal of Electronics and Computer Engg., Spring,
2005.
 Technique Used: Comparative performance analysis of
PCA with Gabor Wavelets and Discrete Gabor
Wavelets.
 Conclusion: Gabor Wavelets showed the best
performance on FERET database, as Gabor Wavelets is
least affected by illumination levels.

13
PROBLEM STATEMENT
 Maximum works were proposed the Face Recognition system in
particular case (Either Pose or Illumination or specific Face
Database).
 Till date there is not single face recognition system for fulfilling
the all (or Maximum factors) in real time application.
 Every method have its pros and cons.
 The presenter motivated by the unsatisfactory scenario of Face
Recognition system to enhance the performance of it.

14
AVAILABLE FACE RECOGNITION ALGORITHMS
BASED ON [4]
 Principle Component Analysis (PCA).
 Normalization of Histogram Analysis (NHA).
 Independent Component Analysis (ICA).
 Normalized Cross Correlation (NCC).
 Sum of Absolute Difference (SAD).
 Linear Discernment Analysis (LDA).
 Discrete Wavelets Transform (DWT).
 Gabor Wavelet Transform(GWT).
 Multilayer Appearance-Tensor based (MAT).
 Multiple Descriptor Key point (MDK-SDK).(Partial Face Also) etc.

15
CHALLENGE FOR FACE RECOGNITION [1]
External Occlusion
By other face
Self Occlusion
By non frontal pose
Facial Accessories
By Sunglass
Not Proper
Illumination
Sensor saturation
By under exposure or
over exposure
Limited Field of View
(FOV) By out of camera.

STANDARD FACE RECOGNITION DATABASE [4]
 The Choice of appropriate database to be used based
on targeted work.
 Color FERET Database
 Yale Face Database.
 PIE Database.
 FIA Video Database.
 CBCL Face Recognition Database.
 Expression Image Database.
 Mugs hot Identification Database.
 Identification Database.
 Indian Face Database.
 Face Recognition Data, University of Essex, UK 16

17
CONCLUSIONS AND FUTURE WORK
 The presenter in initial phase of project work, so exact algorithms
yet to be finalized.
 The presenter aware to follow any based paper should have same
face database.
 Modify the any previous mentioned face recognition algorithms to
enhance the recognition rate of identification.
 The proposed work may also target for fusion between two
algorithms.
 When available algorithms will modified or develop, the result will
be compare with based paper or predecessor method result .

18
REFERENCES
[1] Li, Liao and Jain, “Partial Face Recognition, Alignment free Approach”, IEEE
Transaction on Pattern Analysis and Machine Intelligence, VOL 35, No 5, May,
2013.
[2] Mersico, Nappi and Wechsler, “Robust Face Recognition fro Uncontrolled Pose
and Illumination ”, IEEE Transaction on Systems, man and Cybernetics:
systems, VOL. 43, NO. 1, January, 2013.
[3] Park and Savvides , “Individual Kernal Tensor Subspaces for Robust Face
Recognition: A Computationally Efficient Tensor Framework without requiring
Mode Factorization ”, IEEE on Systems, man and Cybernetics: systems, VOL. 37,
NO. 5 , Oct, 2007.
[4] Karim, Lipu, Rahman and Sultana ,“Face Recognition using PCA based
Method”, IEEE, 2010.
[5] Meade, Kumar and Phillips, “Comparative performance of Principle
Component Analysis , Gabor Wavelets and Discrete Gabor Wavelets”, Canadian
Journal of Electronics and Computer Engg., VOL. 30, NO. 2, Spring, 2005.

REFERENCES CONTD..
[6]Kar, Debbarma, Saha and Pal, "Study of Implementing Automated
Attendance System using Face Recognition Technique” International Journal of
Computer and Communication Engineering, VOL. 1, No. 2, July 2012.
[7] Balcoh, Yousaf, Waqar and Baig,”Algorithm for Efficient Attendance
Management: Face Recognition based Approach”, IJCSI (Online), Vol.9, No.1,
July 2012.
[8] Jiang, Sadka and Crooks, "Technical Correspondence-Multimodal Biometric
Human Recognition for Perceptual Human-Computer Interaction ” IEEE
Transaction on Systems, man and Cybernetics-Part C:Applications and Review,
VOL. 40, NO. 6, November, 2010.
[9]Jyoti, Chadha, Vaidya and Roja,”A robust, low-cost approach to Face
Detection and Face Recognition”, CiiT International Journal of Digital Image
Processing, ISSN 0974-9586(Online), Vol. 15, No 10, October 2011.
[10] Lu and Tan,”Cost-Sensitive Subspace Analysis and Extensions For Face
Recognition”,IEE transactions on Information forensics and Security, Vol. 8, No
3, March 2013.
19

REFERENCES CONTD..
[11] Toole, Philips, Jiang and Abdi,”Face Recognition Algorithms Surpass
Human Matching Faces over changes in Illumination”, IEEE
Transactions on Pattern Analysis and Machine Intelligence, VOL. 29, No.
9, September 2007.
[12]Zhang, Shan, Chen and Gao,”Local Gabor Binary Patterns Based on
Kullback-Leibler Divergence for Partially Occluded Face Recognition”,
IEEE Signal Processing Letters, Vol. 14, No.11, November 2007.
[13] Liu and Liu, “A hybrid Color and Frequency Features Method for
Face Recognition”,IEEE transactions on Image Processing, Vol. 17, No.
10. October 2008.
[14] Mohanty, Sarkar, Kasturi and Phillips, "Subspace Approximation of
Face Recognition Algorithms: An Empirical Study”, IEEE Transactions
on Information Forensics and Security”, Vol. 3, No. 4, December 2008.
[15] “Police use Facial recognition Technology to Nab Rioters”,
http://www.msnbc.msn.com/id/44110353/ns
20

21
Any Question ?
Thank You Watching This PPT.
For Academic Graduates / Post- Graduates/
PhD Scholars Presentation Design
Please Contact
kalyan.acharjya@gmail.com

The PPT is Brought you By
22
KalyanAcharjya,India

Face recognition Face Identification

More Related Content

What's hot (20)

Viewers also liked (14)

Similar to Face recognition Face Identification (20)

More from Kalyan Acharjya (20)

Recently uploaded (20)

Face recognition Face Identification