Because existing face recognition systems lack required accuracy when viewpoint, illumination, expression, occlusion, accessories and so on vary considerably, continued research in this field of biometrics is a challenging objective. It is widely accepted that local features in face images are more robust against such distortions and a spatial-frequency analysis is often more desirable to extract such features. Having good characteristics of space-frequency localization, wavelet analysis is the right choice for this purpose. Among various wavelet bases Gabor functions provide optimized resolution in both spatial and frequency domains. However, only the magnitudes of the Gabor coefficients are used as features for face recognition in most previous works, while the Gabor phases are deemed useless and ignored. But recently researchers are finding impressive results using Gabor phase based feature representations. At the earlier part of this project, we study the characteristics of the phase part of the Gabor wavelet by investigating on 70 different Gabor phase based feature extraction techniques for face recognition. Later, based on the observation, we propose a novel face recognition system using the discriminating capability of the basic 40 Gabor filters. Lastly, we propose two online systems for face verification and identification based on local and global threshold of the similarity measure. Experimental results show that, among the 40 basic Gabor-phase based feature representations, filter responses acquired from larger scales show higher discriminating ability for face recognition. Moreover, our proposed weighted vote based face recognition system shows high accuracy of 98.3% for 1000 subjects for a combination of the FERET, Indian and in-house databases on all three different similarity measures, outperforming the recognition accuracy 83.2% of conventional Gabor phase representation. Finally, the proposed face authentication method based on the 40 basic Gabor phase feature representation and the summation of the 40 basic Gabor phase feature representation show promising results, for images taken under unconstrained pose, illumination and expression changes. As a whole, this project studies different aspects of Gabor wavelet based human face recognition and proposes some novel methods for improved face classification.
We focus on small device level implementation of the algorithm. Such as, face Authencation on Android system based Tablets.
The Face Recognition Framework
Iqbal Nouyed, B. Poon, M. A. Amin, H. Yan "Face Recognition Accuracy of Gabor Phase Representations at Different Scales and Orientations" International Conference on Machine Learning and Cybernetics, Guilin, China 2011 [PDF] [IEEE Xplore][Slides]
The classification capability of 40 different basic Gabor phase representations for human face recognition is examined in this paper. The results show that, among all 40 basic Gabor phase based feature representations, the filter responses acquired from the larger scales show the higher discriminating ability for face recognition.
Iqbal Nouyed, M. Ashraful Amin, Bruce Poon, Hong Yan, "Human Face Recognition Using Weighted Vote of Gabor Magnitude Filters", 7th International Conference on Information Technology and Application, Sydney, Australia 2011 [PDF][IEEE Xplore][Slides]
Typically to extract the Gabor features from facial images, the magnitudes of the Gabor filter responses for different orientations and scales are used. We propose a weighted voting
method using the 40 different Gabor magnitude representations which has shown 100% accuracy using twofold cross validation test where accuracy of conventional method was found to be 95%
for 100 subjects using histogram intersection as similarity measure.
M Ashraful Amin, Iqbal Nouyed, Bruce Poon, Hong Yan "Face Recognition Using Weighted Voting on Gabor Filters", 21st International Conference on Pattern Recognition, Tsukuba Science City, Japan 2012
In this work we present a faster face recognition system implemented using local Gabor binary pattern histogram sequence. The proposed method examines discriminating characteristics of 140 different Gabor face representations acquired using concatenation and summation of filters. Numerical weights are calculated from each filters discriminating ability. Experimental results on FERET and in-house database show that the proposed method can recognize faces with 98.3% accuracy
Papers under Progress:
Discriminating Characteristics of Gabor Phase-Face and an Improved Face Recognition System (Journal Paper)