Presentation of IEEE paper "A Proposed Framework for Robust Face Identification System"
A Proposed Framework for Robust Face Identification System
A Proposed Framework for
Ahmed F. Gad
Faculty of Computers and Information-Menofia University
Face Skin Detection (FSD)
Facial Features Extraction and Enhancement
Algorithms Involved the Proposed System
Databases Used and Results
References and Conclusion
High Level Description
A robust face identification framework to find whether
two face images contains represent the same
Proposed System Phases
System has 4 major phases:
A color constancy algorithm that use the Fast
Fourier Transform and based on RGB.
Pixel values are modified to give an estimation of
original color and ignoring effects.
Image Preprocessing-Retinex Algorithm
Images processed to eliminate degradations and to
be invariant to illumination conditions.
Results are suitable to Human Visual System
To reduce search space for facial features, skin
detection is applied.
Rather than searching whole image for features, only
selected regions are tested.
Which Color Model To Use !
A survey to compare robustness of different color
spaces to detect skin made.
Skin Detection-RGB-H-CbCr Model
A robust color model required to detect skin color
under large scale with high accuracy.
Combination from three color models RGB, HSV,
and YCbCr is used.
A robust object detector based on cascade of simple
Integral image is a an image representation to
facilitate feature detection.
Examples of features if the
But there are more than 18,000 feature.
Just expressive features are selected to minimize
computation time using the Adaptive Boosting
More than one classifier are used to detect objects.
Algorithm can process 15 frames per second.
Feature Extraction-Viola~Jones Algorithm
Object segmentation is applied on the skin binary
Each extracted object is applied to the algorithm to
extract facial features.
Face regions only will continue.
Accurate Eye Detection
A single eye has three regions:
Dark pupil in the eye center.
Lighter sclera region surrounding the pupil.
Skin surrounding both pupil and sclera.
HSV is a robust color space based on its saturation
channel that can differentiate among the top two
Eye Pupils Detection
Eye centers can be found by analyzing the eye
regions for the darkest area using luminance-
chrominance model such as HSV.
Frangi filter is an accurate visualization and
quantiﬁcation of the human vasculature that tries to
use human vessel geometrical structures such
as Hessian matrix to segment vessel regions.
Eigenvalues are extracted from the Hessian matrix
to find such features.
Mouth and Nose Enhancement-Frangi Filter
Native Viola~Jones algorithm can enlarges the
mouth and nose region boundary.
Frangi filter can efficiently detect their boundaries.
The feature vector with 11 metrics gathered from 12
distances is used to compare the two images.
The Center for Vital Longevity Face Database is
used and system is tested against 2o persons each
having 40 different images and yield an accuracy of
Computation time is 1.56 seconds.
Morel, Jean-Michel, Ana B. Petro, and Catalina Sbert. "Fast implementation of color constancy
algorithms." IS&T/SPIE Electronic Imaging. International Society for Optics and Photonics, 2009.
Oliveira, V. A., and A. Conci. "Skin Detection using HSV color space" H. Pedrini, & J. Marques de
Carvalho, Workshops of Sibgrapi. 2009.pp:1-2
Kaur, Amanpreet, and Bv Kranthi. " Comparison between YCbCr Color Space and CIELab Color
Space for Skin Color Segmentation." International Journal of Applied Info. 3.4 (2012): 30-33.
Ma, Zhanyu, and Arne Leijon. "Human skin color detection in RGB space with Bayesian estimation
of beta mixture models." 18th European Signal Processing Conference (EUSIPCO-2010). 2010.
Phung, Son Lam, Abdesselam Bouzerdoum, and Douglas Chai. "A novel skin color model in ycbcr
color space and its application to human face detection." Proceedings of International Conference
on Image Processing. 2002. Vol. 1. IEEE, pp. I-289.
Bin Abdul Rahman, Nusirwan Anwar, Kit Chong Wei, and John See. "RGB-H-CbCr Skin Colour
Model for Human Face Detection." Faculty of Information Technology, Multimedia University
Viola, Paul, and Michael Jones. "Rapid object detection using a boosted cascade of simple
features." Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and
Pattern Recognition, 2001. CVPR 2001. Vol. 1. IEEE, 2001.
Bob Zhang, Lin Zhang, Lei Zhang, Fakhri Karray “Retinal vessel extraction by matched
ﬁlter with ﬁrst-order derivative of Gaussian”, Computers in biology and medicine 40.4
Minear, M. & Park, D.C. “A lifespan database of adult facial stimuli.” Behavior Research
Methods, Instruments, & Computers. 36, .(2004), 630-633.
Dewi Agushinta R, Adang Suhendra, Sarifuddin Madenda, Suryadi H.S. "Face
Component Extraction Using Segmentation Method On Face Recognition System."
Journal of Emerging Trends in Computing and Information Sciences 2.2 (2011). 67-72
Kim, Hyun-Chul, Sung Yang Bang , Sang-Youn Lee. "Face recognition using the
second-order mixture-of-eigenfaces method." Pattern Recognition 37.2 (2004): 337-
Ma, Bingpeng, Yu Su, and Frédéric Jurie. "Covariance descriptor based on bio-inspired features
for person re-identification and face verification." Image and Vision Computing 32.6 (2014):