face recognition using new image representations

27
Face Recognition Using New Image Representations Zhiming Liu and Qingchuan Tao 2009 IEEE

Upload: oro

Post on 22-Feb-2016

25 views

Category:

Documents


0 download

DESCRIPTION

Face Recognition Using New Image Representations. Zhiming Liu and Qingchuan Tao 2009 IEEE. Outline. Introduction Motivation New Image Representation Via PCA Transformation Experiments Conclusion. Introduction. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Face Recognition Using  New Image Representations

Face Recognition Using New Image RepresentationsZhiming Liu and Qingchuan Tao2009 IEEE

Page 2: Face Recognition Using  New Image Representations

Introduction Motivation New Image Representation Via PCA

Transformation Experiments Conclusion

Outline

Page 3: Face Recognition Using  New Image Representations

While the commonly used gray-scale image is derived from the linear combination of R, G, and B color component images, the new Image representations are derived from the Principal Component Analysis (PCA) tranform upon the hybrid configurations of different color component images.

Introduction

Page 4: Face Recognition Using  New Image Representations

We propose to encode the facial information from the new image representations by using an effective Local Binary Pattern (LBP) feature extraction method, which extracts and fuses the multi-resolution LBP features.

Introduction

Page 5: Face Recognition Using  New Image Representations

For color face image recognition, the RGB color space is commonly used in some methods.

As YIQ, HSV, and YCbCr transformed from the RGB space, are adopted to perform face recognition.

Motivation

Page 6: Face Recognition Using  New Image Representations

First, we calculate the correlation coefficients contained between the individual components in RGB, YIQ, and YCbCr color spaces.

Motivation

Page 7: Face Recognition Using  New Image Representations

Based on the within-class scatter matrix Sw and the between-class scatter matrix Sb of the training database, we can evaluate the class separability by using the Fisher criterion: J4 = tr(Sb)/tr(Sw).

Motivation

Page 8: Face Recognition Using  New Image Representations

Sw:類別內散佈矩陣 (within-class scatter matrix )

Sb:類別間散佈矩陣 (between-class scatter matrix )

Motivation

Page 9: Face Recognition Using  New Image Representations

Table II gives the calculation results, which indicate that the color components G and B have the weakest power of image classification, at least for the FRGC training database.

Motivation

Page 10: Face Recognition Using  New Image Representations

We assume that , , and are coloumn vectors: where N=mxn.

We can form a data matrix using all the training images:

where l is the number of training images.

New Image Representation Via PCA

Page 11: Face Recognition Using  New Image Representations

The covariance matrix of may be formulated as follows :

where is the expectation operator, t denotes the transpose operation, and .

New Image Representation Via PCA

Page 12: Face Recognition Using  New Image Representations

The PCA of a random vector X factorizes the covariance matrix into the following form:

where is an orthonormal eigenvector matrix and

is a diagonal eigenvalue matrix with diagonal elements in decreasing order .

New Image Representation Via PCA

Page 13: Face Recognition Using  New Image Representations

Then a new image representation can be derived by projecting three color component images of an image onto :

New Image Representation Via PCA

Page 14: Face Recognition Using  New Image Representations

In particular, the training set contains 12,776 images that are either controlled or uncontrolled.

The target set has 16,028 controlled images and the query set has 8,014 uncontrolled images.

Experiments

Page 15: Face Recognition Using  New Image Representations

A. Effectiveness of New Image Representations for Face Recognition

Some new image representations, such as URCrQ , URCbQ , and so on, can be generated by using the transformation derived from PCA.

Note that before transformation, in (4) are normalized to have zero mean and unit variance, respectively.

Experiments

Page 16: Face Recognition Using  New Image Representations

Table III shows the face verification rates (FVR) at 0.1% false accept rate (FAR) , where only image representations with FVR beyond 60% are listed, and R, Y, and URGB are also included for comparison.

Experiments

Page 17: Face Recognition Using  New Image Representations

Fig. 1 shows some color component images and the resulting new image representations by using the transform coefficients.

Experiments

Page 18: Face Recognition Using  New Image Representations

Table IV show that there are strong decorrelations between UYCbQ and UYCrQ, URCrQ.

Experiments

Page 19: Face Recognition Using  New Image Representations

The fused classification results are detailed in Table V, which indicates that the best performance 77.10% , can be reached by fusing UYCrQ and UYCbQ, as expected.

Experiments

Page 20: Face Recognition Using  New Image Representations

B. LBP-based Face Recognition Using New Image Representation

In this section, we present an effective method to use LBP features for face recognition.

The LBP operator is defined as follows:

Experiments

Page 21: Face Recognition Using  New Image Representations

After extensions, LBP can be expressed as: , where P and R mean P sampling

points on a circle of radius R.

A LBP multi-resolution feature fusion is proposed as shown in Fig. 2.

Experiments

Page 22: Face Recognition Using  New Image Representations

The third set of experiments evaluates face recognition performance by using the proposed multi-resolution LBP feature fusion on new image representations.

Experiments

Page 23: Face Recognition Using  New Image Representations

The proposed LBP method is implemented to UYCrQ, UYCbQ, R, and Y images, and the corresponding experimental results are shown in Table VI.

Experiments

Page 24: Face Recognition Using  New Image Representations

The final results are given in Table VII, which indicates that the best FVR of 83.41% at 0.1% FAR is achieved by fusing the classification outputs of UYCrQ and Y images.

Experiments

Page 25: Face Recognition Using  New Image Representations

Fig . 3 shows the corresponding ROC curves for the best FVR obtained by our method.

Experiments

Page 26: Face Recognition Using  New Image Representations

The experiments show the satisfactory results have been achieved by using these new images and LBP features.

The future work will be focused on seeking the more reliable criteria to choose the color component images, as well as the new learning methods to derive the color transformation.

Conclusion

Page 27: Face Recognition Using  New Image Representations

Thank you for your listening