DO I KNOW YOU? OPTIMIZING THE FACIAL RECOGNITION SYSTEM THROUGH - PowerPoint PPT Presentation

DO I KNOW YOU? OPTIMIZING THE FACIAL RECOGNITION SYSTEM THROUGH DISTANCE METRICS KEVIN PARK UNCW REU: STATISTICAL LEARNING AND DATA MINING 7/25/2017

2 ACKNOWLEDGEMENTS  With all my heart, I express sincere gratitude and thankfulness to my REU family.  Thank You to National Science Foundation and University of North Carolina Wilmington for sponsoring this summer experience.

3 1 2 3 I. II. III. Introduction Experimental Results and Design Conclusion OUTLINE

1. INTRODUCTION

5 WHAT IS FACE RECOGNITION?  Mechanism for a computer to determine whether a given face image is identifiable or unknown  Use machine learning and computer vision techniques to match new, unseen face images to a given set of trained images computer has already seen Known  Face Recognition has been developing interest Image with its wide applications in security, biometrics, and industry Unknown Source: http://www.nec.com/en/global/solutions/safety/face_recognition/images/Face_Recognition_FR_Pic.png

6 PROJECTION ALGORITHMS AND DIMENSION REDUCTION Fisherfaces Eigenfaces Image (Principal Component Analysis) (Linear Discriminant Analysis) “Transform a face image into a set of LDA obtains discriminative information by   crucial features, called eigenfaces” (Turk) maximizing the ratio between group and within Convert group variance which optimizes class separability Image to These eigenfaces are put into an  Vector eigenvector which contains features that Face space is then obtained after this maximization  explain the variation between face images Too Many Dimensions!! Dimension Reduction Source: Turk, M.a., and A.p. Pentland. "Face Recognition Using Eigenfaces." IEEE Computer Society Conference on Computer Vision and Pattern Recognition (1991): 586-91. Web. Source: http://courses.ee.sun.ac.za/Pattern_Recognition_813/lectures/lecture01/img33.png

7 DISTANCE METRICS Name Formula Name Formula Image One Image Two Euclidean Bray Curtis City Block Canberra How to Compare Cosine Correlation Them? Distance Mahalanobis Chebyshev Cosine d = D i = max | x i – y i | [(x − y) t S − 1 (x − y)] 1/2 Distance Metrics

8 RESEARCH PROJECT GOALS  Objective:  1) Determine which combination of dimension reduction techniques and distance metrics will provide the greatest accuracy in the facial recognition task.  2) Assess the impact of combining multiple distance metrics (decision fusion-like) on face recognition accuracy. Distance Distance Distance Distance Distance Metric One Metric T wo Metric Three Metric Four Metric Five Eigenface Fisherface

1I. EXPERIMENTAL DESIGN

10 DATA USED Sample Images  The Morph II Database is a collection of 55,134 face images of 13,617 unique subjects.  The 55,134 mugshots are from 2003 and 2008 and include images of individuals that were arrested once or multiple times. Cleaned Images  The size and longitudinal features of Morph II make it a widely used data source for computer vision and machine learning research.

11 SUBSET SCHEME Subset The subset was developed to only contain • individuals with 10 or more face images in Morph II. Morph II 166 subjects with a total of 1,660 images • • 83 males and 83 females. 126 black individuals, 39 white subjects, and • 1 Hispanic person. • The age distribution of the subset ranges from 16 to 61 years of age where 75% of the subjects are 41 years of age or younger.

12 EXPERIMENTAL SETUP For every subject, nine out of the ten face  Leave One Image Out Setup images are randomly assigned to training. The tenth image is then assigned to testing. Facial recognition accuracy is calculated once the test images are matched to a training image.

13 RESEARCH PROJECT PROCEDURE: FACIAL RECOGNITION 1,494 Training Images and 166 Testing Images 100 principal components Euclidean Distance City Block Distance Eigenface (PCA) Cosine Distance Correlation Distance Classification Image Facial Bray Curtis Distance Method Recognition - Nearest Canberra Distance Neighbors Fisherface (LDA) Chebyshev Distance M. Cosine Distance Projection Distance Metrics Algorithms

14 IMPORTANCE OF DISTANCE METRICS  A new face image will be matched to an image by comparing the test image’s feature vector to every training image’s feature vector.  “Distances between two vectors can be treated as a measure of dissimilarity of the two biometric samples” (Yassin) Training Feature Vector 1 Training Feature Vector 2 Facial Calculate Testing Feature Vector Recognition Distance Classification Metric Comparison Training Feature Vector (n)

15 DECISION FUSION IDEA USING DISTANCE METRICS Proposed Method Method Used Before Test Train Train Test Image Image Image Image Calculate distance with Distance Distance Distance single distance metric Metric N Metric 1 Metric 2 Using Apply Apply Re-Scale minimum Nearest Nearest Distance distance Neighbors Neighbors Facial Recognition Facial Recognition Clasification Clasification

1II. RESULTS AND CONCLUSION

17 DISTANCE METRICS RESULTS Experiment Attributes: • Chebyshev distance metric resulted in the lowest • Leave One Out: 9 Images or Training and1Image for Testing accuracy for face recognition. • 1,494 Training Images and 166 Testing Images • Face Recognition Accuracy under • 100 Principal Components Eigenfaces: 69% to 78% Fisherfaces: 75% to 95%. Euclidean City Cosine Correlation Bray Canberra Chebyshev Mahalanobis Cosine Block Curtis Eigenfaces 69.27% 75.90% 68.07% 70.48% 78.31% 77.71% 46.98% 78.91% Fisherfaces 89.15% 81.92% 95.18% 94.57% 93.97% 75.90% 75.90% 80.72%

18 DECISION FUSION PRELIMINARY RESULTS Euclidean City Cosine Correlation Bray Canberra Chebyshev Mahalanobis Cosine Block Curtis Eigenfaces 69.27% 75.90% 68.07% 70.48% 78.31% 77.71% 46.98% 78.91%

19 CONCLUSION AND FURTHER DIRECTIONS  The experiments illustrate that a particular distance metric with a corresponding projection technique can play a major role in increasing face recognition accuracy.  As the Cosine and Mahalanobis Cosine distance were the best distance metrics, there seems to be evidence to suggest that studying the angle and orientation between two feature vectors is a crucial component in facial recognition.  Studying the orientation rather than physical distance may be a more promising direction to explore in face recognition.  An interesting application to further explore would involve using distance metrics with elastic bunch graphing method in face recognition.  In regards to decision fusion, future work can try more involved weighting schemes and different combinations of distance metrics.

20 REFERENCES  Liu, Kuan-Hsien, ShuichengYan, and C.-C. Jay Kuo. "Age Estimation via Grouping and Decision Fusion." IEEE Transactions on Information Forensics and Security 10.11 (2015): 2408-423. Web.  Seo, Naotoshi. "Eigenfaces and Fisherfaces." ENEE633 Pattern Recognition , Jan. 2017. University of Maryland .  Turk, M.a., and A.p. Pentland. "Face Recognition Using Eigenfaces." IEEE Computer Society Conference on Computer Vision and Pattern Recognition (1991): 586-91. Web.  Yassin, Dk H. Phm, S. Hoque, and F. Deravi. "Age Sensitivity of Face Recognition Algorithms." 2013 Fourth International Conference on Emerging Security Technologies (2013): n. pag. Web.