Sunteți pe pagina 1din 8

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/318348766

A Study on Face Recognition Techniques with Age and Gender Classification

Conference Paper · May 2017


DOI: 10.1109/CCAA.2017.8229960

CITATIONS READS

2 1,197

3 authors, including:

Sandeep Kumar Jagdish Kumar


PEC University of Technology PEC University of Technology
5 PUBLICATIONS   7 CITATIONS    65 PUBLICATIONS   212 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

nonlinear control system View project

Model Reference adaptive control using Artificial Neural network View project

All content following this page was uploaded by Sandeep Kumar on 11 July 2017.

The user has requested enhancement of the downloaded file.


A Study on Face Recognition Techniques with Age
and Gender Classification
Sandeep Kumar Sukhwinder Singh Jagdish Kumar
Department of Electronics and Department of Electronics and Department of Electrical
Communication Engineering Communication Engineering Engineering
PEC University of Technology, PEC University of Technology, PEC University of Technology,
Chandigarh-160012, India Chandigarh-160012, India Chandigarh-160012, India
sandeepkumar.phdece15@pec.edu.in sukhwindersingh@pec.ac.in jagdishkumar@pec.ac.in

Abstract- Nowadays research has explored to extracting shown in figure 1. From these attributes biometric system can
auxiliary information from various biometric techniques such as extract some information from the data and compares this
fingerprints, face, iris, palm, voice etc. This information contains information with the templates in the database to determine an
some features like gender, age, beard, moustache, scars, height, identity.
hair, skin color, glasses, weight, facial marks, tattoos etc. All this
information contributes more and more during identification.
The major changes that come across face recognition is to find Face
age & gender of the person. This paper contributes a significant
survey of various face recognition techniques for finding the age
and gender. The existing techniques are discussed based on their Gender Contact Lens Beard
performances. This paper also provides future directions for
further research. AGE Glasses Moustache
Keywords- Biometrics, Age, Face Recognition, Gender.
Hair Cap Makeup
I. INTRODUCTION
Skin Accessories Tattoos
Biometrics is the unique measurable characteristics of
automatically recognizing an individual identification based Scars Facial Marks Moles
on their physical and behavioral characteristics [1].
Physiological are regarding the physique of the body such as Fig. 1: Auxiliary information from the face
fingerprint, face, palm, iris and DNA [2]. On the other hand
behavioral involves the behavior of a person such as signature, II. BACKGROUND
keystrokes and voice.
In the last 51 years, the face recognition system have been In 1960 Woody et al. gave the first semi-automated facial
researched as their low cost and are reliable in the different recognition technique. This technique required some features
fields such as forensic, commercial, health-care, such as mouth, nose, eyes and nose on the image [2][3]. After
entertainment, government, traveling and immigration etc [3]. that, face recognition techniques came in the picture. In 2001
Nowadays the biometric system is a big business in the world. onwards this biometric technique increased rapidly according
The biometrics market grew in the year of 2000 with $400 to figure 2. The main goal of this literature survey is the
million and in the year of 2011biometrics business is $5 justification of the constantly increasing scientific interest in
billion. It is expected that the biometrics business would face recognition biometric technique. The literature analysis
increase up to $23 billion in the year of 2019 [10]. Due to their can be achieved by analyzing the publications of the last 51
intrusive property success of some biometrics systems (i.e. years related to the face. This complete database is used from
fingerprint, palm and iris) depends on user cooperation but well known Scopus database [4]. This database includes
face recognition systems are non-intrusive as it doesn‟t require enough information for our analysis, hence it is accepted by
physical interaction of users to be identified once the input is the research and scientific community. This analysis has been
acquired [5]. One more important characteristic is Passive performed by applying the keyword „face recognition‟. The
Identification which means this biometric system identifies a database of face recognition for various types of publications
person in one-to-many environments (e.g. person in a crowed namely, Article, Journals, Review, Book Chapters, Conference
place) [6]. This is why personal authentication face Papers and Book as shown in figure 3. Moreover, the
recognition is considered as the friendly among all biometrics. importance of face recognition is justified by the publication
The face biometric system can identify various attributes from of more articles and conference paper. According to figure 3.
the image such as gender, age, beard, moustache, scars, height, number of book and book chapter is quite small so there is a
hair, skin color, glasses, weight, facial marks, tattoos etc as scope of improvement in face recognition techniques. In the
Fig. 2: Face Recognition related publications for the past 51 years

history of face recognition the last 15 years have been found to identification [8] [9]. Facial segmentation accuracy depends
be most productive years in which maximum papers have been on the pose, noise, lighting conditions and distance between
published. This data is very high which means that the the object and the camera. The various types of challenges
research in the field of face recognition experiences its highest coming in the picture during face detection are described
evolution so far. The outcomes of this analysis should be below:
translated to more research activities and high prior-
A. Pose
knowledge must favor the discovering and developing of the
next generation frameworks in face recognition and The most challenging situation is that the human face
applications. varies with respect to the relative camera-face pose (45
degrees, profile, frontal and upside down).
B. Facial Expression
The facial expression such as anger, fear, disgust,
happiness, sadness and surprise is most influential
temperaments for human beings to communicate their
feelings.
C. Illumination
Illumination is a major challenge during face recognition.
This factor is related to the lightening and angle of the light.
D. Occlusion
Occlusion is the main challenge during face recognition
because sometimes the face is partially covered and occluded
by others objects.
E. ImagingCondition
Fig. 3: Face related publications for the past 51 years for different Documents During the face image capture some factors such as
different lightening conditions and camera characteristic
III. CHALLENGES of FACE RECOGNITION (lenses, sensor response) affect the face recognition accuracy.
Most of the applications such as computer, human crowd F. Different Facial Features
surveillance, face processing, artificial intelligent, content
Different type of facial features such as glasses, beard, hair
based image retrieval and video surveillance etc. require face
moustache, scars, moles, tattoos, skin colors and makeup
detection for identification and verification the enrolled users
affect the face recognition accuracy.
[7]. The skin color segmentation is the major problem for face
G. Face Size features, skin color, shape and texture are used [14, 15]. These
This factor is also a major challenge because face size can methods are very sensitive to illumination, occlusion,
vary a lot person to person. Not only different people have existence of skin color regions, and adjacent faces as shown in
different sized faces but the face closer to the camera and far figure 5.
away from the camera also pose a challenge.
H. Facial Expression
This factor depends on the nature of the images i.e. face
images vary at different camera rotations about its optical axis.
IV. CHALLENGES of FACE RECOGNITION
Face recognition is the sub-area of pattern recognition
research and technology. Overview of face recognition system
is shown in figure 4. Firstly we take input image from the (a) (b)
standard cameras and this image is known as input image [3].
Secondly pre-processing is performed to improve the image
quality and to reduce the noise in the taken images. Next
phase is face segmentation where face part is a cropped from
the human body and background of the image. In the next
phase, feature extraction is done after segmenting the input
image to a very good level and relevant features are identified
which can help in distinguishing that person to other people.
In the last phase, a template is generated for enrollment and (c) (d)
matching purpose. Finally, a matching is done at the biometric
system for identification of the authenticated person, if there Fig 5: (a) Original Image, (b) Skin Color Binary Image, (c) Skin Color
is a match then output will come in the form of user accepted Mark Area, (d) Image verified by face length breadth ratio [14]
otherwise user rejected [2][7]. Face recognition systems can C. Template-based methods
be divided into different categories: Template based methods are sensitive to pose, scale and
A. Knowledge-based methods shape variation of the human body. Deformable template
The knowledge-based techniques are based on geometry of methods have been proposed to deal with such variations of
the face and arrangement of the facial features [11]. These pose, scale and shape of the body. Template based methods
knowledge based methods describe the shape, size and texture. using the elastic models include shape parameters as well as
Some methods describe few other characteristics of facial intensity information of facial features.
features such as head, eyebrows, eyes, nose and chin [12]. The
major problem of these techniques is that they do not perform
well because of different types of pose or head orientations as
shown in my implementation results of V-J algorithm in figure
4 [13].

Fig 6: A 14x16 pixel ratio template [8]

D. Template-based methods
The appearance based methods used for face detection are
Eigen faces, Linear Discriminant Analysis, Neural Networks,
Support Vector Machine and Hidden Markov Models. With
the help of these methods the whole image is scanned and
image regions are identified as face or non face. During
template matching appearance based methods will come in the
Fig 4: Face Segmentation Results obtained on CMU dataset images picture.
B. Feature invariant approaches First main step of face recognition system is face
Main aim of these techniques is to find structural features segmentation and comparison of various types of face
of the human face even with lighting conditions varying. detection techniques as shown in Table I. According to
Different types of structural features such as facial local literature survey face detection techniques have low accuracy
so there is a scope of improvement in term of accuracy. Georgia Tech 50 people of images.
Second main step of face recognition system is the feature Indian Movie 34512 images of 100 Indian actors are in different age,
database pose, gender, expression and occlusion.
extraction and according to Table II performance is not JAFFE 213 images of 7 facial expressions each female
optimum so there is a scope of further improvement in feature Labeled 8,500 faces (taken from Wikipedia).
extraction techniques. Various types of face databases which Wikipedia faces
are easily available for research community on the internet and LDHF Database 100 people‟s images at different distances of 60m,
some well known face database are shown in Table III. 100m, and 150m from the subject‟s.
MIT-CBCL 200 images are in different illumination, pose, and
database complex background.
TABLE I: COMPARISON OF EXISTING TECHNIQUES OF FACE YouTube 604 images of 151 subjects
DETECTION MORPH 55,134 images of 13,000 people‟s. This database
Database describes gender, date of birth, and date of acquisition.
Ref Segmentation Dataset Performance 3D face 1149 of facial color of 105 adults.
Techniques Oulu Face 125 different faces are in different illuminant
Database conditions.
[12] RGB Colour Space Self-prepared Accuracy of DR= 80%
Yale face 165 greyscale images of 15 individuals - 11 images per
120 Images
Database subject.
[16] YcbCr + RGB + FRGC 302 CDR = 82.7%
YouTube Faces 3,425 videos of 1,595 different people‟s.
Canny Edge Images FPR = 22.8%
The PUT Face 9971 images of 100 people‟s.
MR = 17.2%
[17] Skin Colour Self-prepared DR= 88.9%
[18] Adaboost & SVM MIT + CMU ER=5.41 Biometrics is the unique measurable characteristics of Despite
(2500 Images) the advantages of iris as a biometric modality, iris spoofing
[19] RGB-H-CbCr Self-prepared FDR = 28.29%
DSR = 90%
poses a threat where an imposter impersonates an enrolled
user‟s identity to gain the user‟s privileges. This poses a great
DR: Detection Rate, CDR: Correct detection Rate, FPR: False Positive Rate, challenge to iris recognition when used in large-scale
MR: Missing Rate, ER: Error Rate,, FDR: False Detection Rate, DSR:
Detection Success Rate applications. The detection of fake iris due to printed iris
image or textured contact lens can be broadly categorized as:
i) recognizing iris textural pattern, ii) exploiting biological
TABLE II: COMPARISON OF EXISTING TECHNIQUES OF FACE characteristics for iris liveness detection and iii) analyzing
FEATURE EXTRACTION
physical properties of the iris.

Ref Segmentation Feature Dataset Performance V. AGE


Techniques Extraction Face recognition systems are essential in real time
[20] Haar Features LBP CK+ and 89.64% and applications such as for identification of the missing children
and Adaboost JAFFE 85.06% and passport verification because there is significant age
[21] Adaboost LBP and FDDB DR = 0.77% difference between investigate and gallery images [26]. One
MCT and MIT and 0.91%
+ CMU more practical application in which age is a condition for
[22] Skin Color ANN LFW and CR = 89.5% access to a product is alcohol or tobacco. Table IV represents
PEAL MDR= 10.5 the comparison of various techniques to find out the ages of
FDR = 0.5%
[23] Viola-Jones SVM Self- Training time
prepared period 60%
300 increases
Images
[24] Hough SVM AR FR rate
Transform database increased by
7.7%
LBP: Local Binary Pattern, ANN: Artificial Neural Network, SVM: Support
Vector Machine, MCT: Modified Census Transform, DR: Detection Rate,
MDR: Missing Detection Rate, FDR: False detection Rate

TABLE III: COMPARISON OF DIFFERENT FACE DATABASE [25]

Database Description
AT & T 40 people with 10 images of each (different lighting
Database conditions). Fig. 7: Age related publications for the past 10 years
AR Database 4,000 color images of 70 men‟s and 56 women‟s.
CAS-PEAL Face 99,594 images of 1040 people are in different Pose, different categories. Two important factors play a vital role in
Database Expression, Accessory, and Lighting conditions. the real life. Firstly, accumulated data suggest that age
CMU Multi-PIE 750,000 images of 337 people.
CK 486 images of 97 people‟s estimation faces can be quite accurate, but there is scope of
FERET 1000 people of images further improvement because age factor is sensitive and
Kinect Face 52 people in different lighting and occlusion depends on the basis of group membership [27]. Secondly, it
Database conditions.
is proved that human face aging depends on ethnicity and preprocessing step for face recognition. The face recognition
genders. This analysis gives right directions for further has been explored in a large manner by many researchers but
research to identify the ages from the images. very few works have been explored in case of gender
TABLE IV
classification [28]. The face image is used for classifying the
EXISTING INTRODUCTORY OVERVIEWS ON AGE ESTIMATION gender, so the gender classification process can make face
TECHNIQUES recognition twice as fast by reducing the search time for
recognizing the person. Table V represents the comparison of
Ref Feature Database Accuracy
various techniques for identification to the gender.
Extraction
TABLE V: EXISTING INTRODUCTORY OVERVIEWS ON GENDER
[31] SVM FG-NET 4.07% ESTIMATION TECHNIQUES
MORPH 5.05 ± 0.11
[32] PCA+USA FG-NET 94.87% Ref. Feature Classifiers Database Accuracy
[33] Deep NN eWIT 65.8% Extraction
[34] NTCA+LBP FG-NET 48.96% [46] LBP NN FERET 80.2%
MORPH 83.8% [47] PCA LDA FERET 85%
[35] PCA/SVM CACD 94.2% [48] LBP and WLD Nearest FERET 98.82%
MORPH 92.8% Neighbor
[36] Gabor/HOG WhoIsIt 28.53 ± 1.03 [49] PCA SVM & K-NN Self- 93%
FG-Net 20.34 ± 0.47 prepared
[37] KDCV / NN FG-NET 93% Database
[38] SVM LFW 88.6 ± 0.3 [50] PCA CDLA FERET 87.44%
[39] AAM FG-NET 95% [51] Adaboost Fusion FRGCv2 86%
[40] GOP+SVM Passport EER=30.5% [52] PCA Euclidean FERET 82%
[41] GB FG-NET 82.7% [53] PCA LDA FRGC 55.47%
[42] RMIR MORFH MAE=8.6 [54] PCA SVM Self- 93.33%
[43] DML FG-NET 84% prepared
[44] SSE FG-NET MAE=5.21 [55] LBP Adaboost Private 96.32%
[45] AMF YGA MAE=8(M), 7.8(F) [56] LBP Adaboost FERET 95.75%
[57] PCA NN Private 88.7%
SVM: Support Vector Machine, PCA: Principal Component Analysis, [58] ICA LDA FERET 99.3%
LBP: Local Binary Pattern, NN: Neural Network, HOG: Histogram of [59] PCA SVM Private 91.1%
Oriented Gradient, AAM: Active Appearance Model
LBP: Local Binary Pattern, PCA: Principal Component Analysis, NN:
Neural Network, LDA: Linear Discriminative Analysis, SVM: Support
VI. GENDER Vector Machine
Gender classification is an easy task for humans to identify
a person according to his or her face but it‟s a very VII. CONCLUSION
challenging task for computers [28]. Last 10 years of gender This paper presents a comprehensive survey of face
classification for different databases are shown in figure 7. recognition and various challenges in the last 51 years. This
This analysis gives fruitful directions for further research in literature analysis showed continuously increasing interest in
the field of gender classification [29]. the field of face recognition. During face recognition some
well known problems such as pose, facial expression,
illumination, occlusion, different facial features etc. have
attained a lot of attention in the research community of
computer vision and pattern recognition. Various types of
techniques have been proposed to compensate to all these
challenges but still there are some unsolved challenges, so
there is a scope of optimization. All these analysis will give a
right direction to the researcher in future to resolve the
unsolved challenges.
REFERENCES
[1] Zhang, Xiaozheng, and Yongsheng Gao. "Face recognition across pose:
A review" Pattern Recognition, vol 42, no. 11, pp. 2876-2896, 2009.
[2] Tolba, A. S., A. H. El-Baz, and A. A. El-Harby. "Face recognition: A
literature review." International Journal of Signal Processing, vol 2, no.
2, pp. 88-103, 2006.
[3] Dantcheva, Antitza, Petros Elia, and Arun Ross. "What else does your
Fig. 8: Gender related publications for the past 10 years biometric data reveal? A survey on soft biometrics." IEEE Transactions
on Information Forensics and Security, vol 11, no. 3, pp. 441-467, 2016.
Gender classification has gained a lot of interest in the
[4] Papakostas, George A. "Over 50 Years of Image Moments and Moment
research community because of its wide application such as Invariants." Moments and Moment Invariants-Theory and Applications,
man-machine communication [30], security, law enforcement, pp. 3-32, 2014.
demographics studies, education and telecommunication, etc
as shown in figure 8. Gender classification is a useful
[5] Zafeiriou, Stefanos, Cha Zhang, and Zhengyou Zhang. "A survey on [27] Rothe, Rasmus, Radu Timofte, and Luc Van Gool. "DEX: Deep
face detection in the wild: past, present and future." Computer Vision EXpectation of apparent age from a single image." IEEE International
and Image Understanding, vol. 138 pp. 1-24, 2015. Conference on Computer Vision Workshops, pp. 10-15, 2015.
[6] Jin, Jing, Bin Xu, Xiaoliang Liu, Yuanqing Wang, Liqun Cao, Lei Han, [28] Bissoon, Terishka, and Serestina Viriri. "Gender classification using
Biye Zhou, and Minggao Li. "A face detection and location method face recognition." The International Conference on Adaptive Science
based on Feature Binding." Signal Processing: Image and Technology, pp. 1-4, 2013.
Communication, vol. 36, pp. 179-189, 2015. [29] Yang, Zhiguang, Ming Li, and Haizhou Ai. "An experimental study on
[7] Hatem, Hiyam, Zou Beiji, and Raed Majeed. "A Survey of Feature Base automatic face gender classification." In 18th International Conference
Methods for Human Face Detection." International Journal of Control on Pattern Recognition (ICPR'06), pp. 1099-1102, 2006.
and Automation, vol. 8, no. 5, pp. 61-78, 2015. [30] Sun, Zehang, George Bebis, Xiaojing Yuan, and Sushil J. Louis.
[8] Rizvi, Qaim Mehdi, Bal Gopal Agarwal, and Rizwan Beg. "A Review "Genetic feature subset selection for gender classification: A comparison
on Face Detection Methods.", 2011. study." In Sixth IEEE Workshop on Applications of Computer Vision,
[9] Yang, Ming-Hsuan, David J. Kriegman, and Narendra Ahuja. "Detecting (WACV 2002), pp. 165-170, 2002.
faces in images: A survey." IEEE Transactions on pattern analysis and [31] Wang, Shengzheng, Dacheng Tao, and Jie Yang. "Relative Attribute
machine intelligence, vol. 24, no. 1, pp. 34-58, 2002. SVM+ Learning for Age Estimation." IEEE transactions on cybernetics,
[10] https://www.bostonglobe.com/ideas/2014/11/23/facial-recognition- vol. 46, no. 3, pp. 827-839, 2016.
technology-goes-way-back/CkWaxzozvFcve Q7kvdLHGI/story.html [32] Li, Zhifeng, Dihong Gong, Xuelong Li, and Dacheng Tao. "Aging Face
[11] Kouzani, Abbas Z., Fangpo He, and Karl Sammut. "Commonsense Recognition: A Hierarchical Learning Model Based on Local Patterns
knowledge-based face detection." IEEE International Conference on Selection." IEEE Transactions on Image Processing, vol. 25, no. 5 pp.
Intelligent Engineering Systems ( INES'97), pp. 215-220, 1997. 2146-2154, 2016.
[12] Devadethan, S., Geevarghese Titus, and S. Purushothaman. "Face [33] Nagpal, Shruti, Maneet Singh, Richa Singh, and Mayank Vatsa.
detection and facial feature extraction based on a fusion of knowledge "Regularized Deep Learning for Face Recognition With Weight
based method and morphological image processing." IEEE Annual Variations." IEEE Access 3, pp. 3010-3018, 2015.
International Conference on Emerging Research Areas: Magnetics, [34] Bouchaffra, Djamel. "Nonlinear topological component analysis:
Machines and Drives (AICERA/iCMMD), pp. 1-5, 2014. application to age-invariant face recognition." IEEE transactions on
[13] Sun, Wei, Weigong Zhang, Xiaorui Zhang, Gang Chen, and Chengxu neural networks and learning systems, vol. 26, no. 7,pp. 1375-1387,
Lv. "Multi-feature driver face detection based on area coincidence 2015.
degree and prior knowledge." The 4th IEEE Conference on Industrial [35] Chen, Bor-Chun, Chu-Song Chen, and Winston H. Hsu. "Face
Electronics and Applications, pp. 222-225, 2009. recognition and retrieval using cross-age reference coding with cross-
[14] Fei, Zhao, and Qiao Qiang. "Face detection based on rectangular age celebrity dataset." IEEE Transactions on Multimedia, vol. 17, no. 6
knowledge rule and face structure." The First International Conference pp. 804-815, 2015.
on Information Science and Engineering, 2009. [36] Singh, Maneet, Shruti Nagpal, Richa Singh, and Mayank Vatsa. "On
[15] Yadav, Shalini, and Neeta Nain. "Fast Face Detection Based on Skin recognizing face images with weight and age variations." IEEE Access 2
Segmentation and Facial Features." The 11th International Conference pp. 822-830, 2014.
on Signal-Image Technology & Internet-Based Systems (SITIS), pp. 663- [37] Ali, Amal Seralkhatem Osman, Vijanth Sagayan, Aamir Malik Saeed,
668, 2015. Hassan Ameen, and Azrina Aziz. "Age-invariant face recognition
[16] Ghimire, Deepak, and Joonwhoan Lee. "A robust face detection method system using combined shape and texture features." IET Biometrics, vol.
based on skin color and edges." Journal of Information Processing 4, no. 2, pp. 98-115, 2015.
Systems, vol. 9, no. 1, pp. 141-156, 2013. [38] Eidinger, Eran, Roee Enbar, and Tal Hassner. "Age and gender
[17] Liu, Qiong, and Guang-zheng Peng. "A robust skin color based face estimation of unfiltered faces." IEEE Transactions on Information
detection algorithm." The 2nd International Asia Conference Forensics and Security, vol. 9, no. 12, pp. 2170-2179, 2014.
on Informatics in Control, Automation and Robotics (CAR), pp. 525- [39] Wang, Yunhong, Zhaoxiang Zhang, Weixin Li, and Fangyuan Jiang.
528, 2010. "Combining tensor space analysis and active appearance models for
[18] Valiollahzadeh, Seyyed Majid, Abolghasem Sayadiyan, and Mohammad aging effect simulation on face images." IEEE Transactions on Systems,
Nazari. "Face detection using adaboosted SVM-Based component Man, and Cybernetics, Part B (Cybernetics), vol. 42, no. 4, pp. 1107-
classifier." arXiv preprint arXiv:0812.2575, 2008. 1118, 2012.
[19] Bin Abdul Rahman, Nusirwan Anwar, Kit Chong Wei, and John See. [40] Ling, Haibin, Stefano Soatto, Narayanan Ramanathan, and David W.
"Rgb-h-cbcr skin colour model for human face detection." Faculty of Jacobs. "Face verification across age progression using discriminative
Information Technology, Multimedia University, 2007. methods."IEEE Transactions on Information Forensics and
security, vol. 5, no. 1, pp. 82-91, 2010.
[20] Happy, S. L., and Aurobinda Routray. "Automatic facial expression
recognition using features of salient facial patches." IEEE transactions [41] J. Suo, S.-C. Zhu, S. Shan, and X. Chen, “A compositional and dynamic
on Affective Computing, vol. 6, no. 1, pp. 1-12, 2015. model for face aging,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 32,
no. 3, pp. 385–401, 2010.
[21] Jeong, Kyungjoong, Jaesik Choi, and Gil-Jin Jang. "Semi-local structure
patterns for robust face detection." IEEE Signal Processing Letters, vol. [42] B. Ni, Z. Song, and S. Yan, “Web image mining towards universal age
22, no. 9, pp. 1400-1403, 2015. estimator,” in Proc. 17th ACM Int. Conf. Multimedia, pp. 85–94, 2009.
[22] Ye, Xueyi, Xueting Chen, Huahua Chen, Yafeng Gu, and Qiuyun Lv. [43] B. Xiao, X. Yang, Y. Xu, and H. Zha, “Learning distance metric for
"Deep learning network for face detection." The 16th International regression by semidefinite programming with application to human age
Conference on Communication Technology (ICCT), pp. 504-509, 2015. estimation,” in Proc. 17th ACM Int. Conf. Multimedia, 2009, pp. 451–
460, 2009.
[23] Mingxing, Jia, Du Junqiang, Cheng Tao, Yang Ning, Jiang Yi, and
Zhang Zhen. "An improved detection algorithm of face with combining [44] S. Yan, H. Wang, Y. Fu, J. Yan, X. Tang, and T. S. Huang,
adaboost and SVM." The 25th Chinese Control and Decision “Synchronized submanifold embedding for person-independent pose
Conference (CCDC), pp. 2459-2463, 2013. estimation and beyond,” IEEE Trans. Image Process., vol. 18, no. 1, pp.
202–210, 2009.
[24] Kawulok, Michal, and Janusz Szymanek. "Precise multi-level face
detector for advanced analysis of facial images." IET image processing, [45] X. Zhuang, X. Zhou, M. Hasegawa-Johnson, and T. Huang, “Face age
vol. 6, no. 2, pp. 95-103, 2012. estimation using patch-based hidden Markov model supervectors,” In
Proc. 19th Int. Conf. Pattern Recognit., pp. 1–4, 2008.
[25] https://www.kairos.com/blog/60-facial-recognition-databases
[46] Amilia, Sindi, Mahmud Dwi Sulistiyo, and Retno Novi Dayawati. "Face
[26] Juefei-Xu, Felix, Khoa Luu, Marios Savvides, Tien D. Bui, and Ching image-based gender recognition using complex-valued neural network."
Y. Suen. "Investigating age invariant face recognition based on The 3rd International Conference on Information and Communication
periocular biometrics." The International Joint Conference on Technology (ICoICT), pp. 201-206, 2015.
Biometrics (IJCB) , pp. 1-7, 2011.
[47] Bissoon, Terishka, and Serestina Viriri. "Gender classification using
face recognition." The International Conference on Adaptive Science
and Technology, pp. 1-4, 2013.
[48] Alamri, Taghreed, Muhammad Hussain, Hatim Aboalsamh, Ghulam
Muhammad, George Bebis, and Anwar M. Mirza. "Category Specific
Face Recognition Based On Gender." The International Conference on
Information Science and Applications (ICISA), pp. 1-4, 2013.
[49] Sakarkaya, Mutlu, Fahrettin Yanbol, and Zeyneb Kurt. "Comparison of
several classification algorithms for gender recognition from face
images." The IEEE 16th International Conference on Intelligent
Engineering Systems (INES), pp. 97-101, 2012.
[50] Chen, Duo, Jun Cheng, and Dacheng Tao. "Clustering-based
discriminative locality alignment for face gender recognition." The IEEE
International Conference on Intelligent Robots and Systems, pp. 4156-
1461, 2012.
[51] Ballihi, Lahoucine, Boulbaba Ben Amor, Mohamed Daoudi, Anuj
Srivastava, and Driss Aboutajdine. "Boosting 3-D-geometric features for
efficient face recognition and gender classification." IEEE Transactions
on Information Forensics and Security, vol. 7, no. 6, pp. 1766-1779,
2012.
[52] Akbari, Rohollah, and Saeed Mozaffari. "Performance enhancement of
PCA-based face recognition system via gender classification method."
The 6th Iranian Conference on Machine Vision and Image Processing,
pp. 1-6, 2010.
[53] Hwang, Wonjun, Haibing Ren, Hyunwoo Kim, Seok-Cheol Kee, and
Junmo Kim. "Face recognition using gender information." The 16th
IEEE International Conference on Image Processing (ICIP), pp. 4129-
4132, 2009.
[54] Zhang, De, and Yun-Hong Wang. "Gender recognition based on fusion
on face and gait information." The International Conference on Machine
Learning and Cybernetics, pp. 62-67, 2008.
[55] Z. Yang and H. Ai, “Demographic classification with local binary
patterns,” In Proc. IAPR Int. Conf. Biometrics, pp. 464–473, 2007.
[56] N. Sun, W. Zheng, C. Sun, C. Zou, and L. Zhao, “Gender classification
based on boosting local binary pattern,” In Proc. 3rd Int. Symp. Adv.
Neural Netw., pp. 194–201, 2006.
[57] A. Khan, A. Majid, and A. M. Mirza, “Combination and optimization of
classifiers in gender classification using genetic rogramming,” Int. J.
Knowl.-Based Intell. Eng. Syst., vol. 9, no. 1, pp. 1–11, 2005.
[58] A. Jain and J. Huang, “Integrating independent components and linear
discriminant analysis for gender classification,” In Proc. 6th IEEE Int.
Conf. Autom. Face Gesture Recognit., pp. 159–163, 2004.
[59] Z. Sun, G. Bebis, X. Yuan, and S. J. Louis, “Genetic feature subset
selection for gender classification: A comparison study,” In Proc. 6th
IEEE Workshop Appl. Comput. Vis., pp. 165–170, 2002.

View publication stats

S-ar putea să vă placă și