Effective Face Recognition Research Articles

Effective and fast face recognition system using hybrid features of orthogonal rotation invariant moments and wavelet transforms

Orthogonal rotation invariant moments (ORIMs) exhibit attractive characteristics such as rotation invariance (translation and scale invariance can be made after normalization process), robustness against noise, good image reconstruction capability, and low information redundancy. Therefore, these are the most commonly used global techniques for image description in different digital image processing applications. The ORIM features are, however, unable to represent local features. Wavelet transforms (WTs), on the other hand, can effectively represent local features. WTs are not invariant to geometric transformation. We propose hybrid face recognition methods based on ORIMs and WTs. In these methods, WTs are used to represent the local features of face images and also to reduce the dimensionality of the face image which results in reducing the large computational time required for moments. WTs retain the characteristics of the original image even after reducing the size of the image. Moreover, the low frequency subband of WTs provides coefficients which are less sensitive to facial expression variation. The role of ORIMs is to capture highly discriminative rotation invariant global features with minimum redundancy. Detail experimental results show that the proposed hybrid methods based on the combination of WTs and ORIMs outperform the ORIMs methods with low processing time.

Prosopagnosia After Stroke: Potentials for Impairment and Treatment

The ability to recognize and identify people and determine how they may be feeling from looking at their faces is an important skill that people normally achieve effortlessly in infancy. Effective face recognition skills remain essential for social competence throughout the life course. A major cause of impairment in face processing, conventionally known as prosopagnosia, is stroke. In this article, the potentials for acquired prosopagnosia after stroke are examined. The incidence of prosopagnosia after stroke is difficult to establish, but in one clinical sample about half of those who survived a right hemisphere stroke had prosopagnosia. The recently published National Clinical Guideline for Stroke 2012 omits reference to assessment for prosopagnosia, which suggests that the personal distress and negative impact on social life that can accompany prosopagnosia is not fully appreciated or at least not considered a priority after stroke. The few published cases where there has been a focused attempt to provide rehabilitation for chronic prosopagnosia suggest that lesions in face-processing areas are resistant to treatment but that some recovery can accompany extended practice. It is concluded that where there is evidence of prosopagnosia following stroke, treatment should be offered, although rehabilitation may be better focused on supporting and extending existing compensatory strategies, such as the use of voice, body shape, and gait to assist in person recognition and, as an important consequence, social functioning.

Mapping Face Recognition Information Use across Cultures

Face recognition is not rooted in a universal eye movement information-gathering strategy. Western observers favor a local facial feature sampling strategy, whereas Eastern observers prefer sampling face information from a global, central fixation strategy. Yet, the precise qualitative (the diagnostic) and quantitative (the amount) information underlying these cultural perceptual biases in face recognition remains undetermined. To this end, we monitored the eye movements of Western and Eastern observers during a face recognition task, with a novel gaze-contingent technique: the Expanding Spotlight. We used 2° Gaussian apertures centered on the observers’ fixations expanding dynamically at a rate of 1° every 25 ms at each fixation – the longer the fixation duration, the larger the aperture size. Identity-specific face information was only displayed within the Gaussian aperture; outside the aperture, an average face template was displayed to facilitate saccade planning. Thus, the Expanding Spotlight simultaneously maps out the facial information span at each fixation location. Data obtained with the Expanding Spotlight technique confirmed that Westerners extract more information from the eye region, whereas Easterners extract more information from the nose region. Interestingly, this quantitative difference was paired with a qualitative disparity. Retinal filters based on spatial-frequency decomposition built from the fixations maps revealed that Westerners used local high-spatial-frequency information sampling, covering all the features critical for effective face recognition (the eyes and the mouth). In contrast, Easterners achieved a similar result by using global low-spatial-frequency information from those facial features. Our data show that the face system flexibly engages into local or global eye movement strategies across cultures, by relying on distinct facial information span and culturally tuned spatially filtered information. Overall, our findings challenge the view of a unique putative process for face recognition.

An Effective Face Recognition in Various Lighting Conditions Using LBP/LTP Techniques

Making recognition is more reliable under uncontrolled lighting conditions is one of the most important challenges for practical face recognition systems. We tackle this by combining following methods. The first step is simple and efficient preprocessing chain. The preprocessing is used to avoid the unwanted illumination effects such as Non-uniform illumination, Shadowing & highlights, aliasing, blurring and noise. Second step includes local binary pattern (LBP) and local ternary pattern methods (LTP). LBP is possible to describe the texture and shape of a digital image. LTP is a generalization of the local binary pattern (LBP). Local texture descriptor that is more discriminant and less sensitive to noise in uniform regions. The final step is used to improve robustness by adding two complementary sources Gabor wavelets and LBP showing that the combination is considerably more accurate than either feature set alone.

An improved discriminative common vectors and support vector machine based face recognition approach

An improved discriminative common vectors and support vector machine based face recognition approach is proposed in this paper. The discriminative common vectors (DCV) algorithm is a recently addressed discriminant method, which shows better face recognition effects than some commonly used linear discriminant algorithms. The DCV is based on a variation of Fisher’s Linear Discriminant Analysis for the small sample size case. However, for multiclass problem, the Fisher criterion is clearly suboptimal. We design an improved discriminative common vector by adjustment for the Fisher criterion that can estimate the within-class and between-class scatter matrices more accurately for classification purposes. Then we employ support vector machine as the classifier due to its higher classification and higher generalization. Testing on two public large face database: ORL and AR database, the experimental results demonstrate that the proposed method is an effective face recognition approach, which outperforms several representative recognition methods.

A novel face recognition approach based on kernel discriminative common vectors (KDCV) feature extraction and RBF neural network

The discriminative common vectors (DCV) algorithm is a recently addressed discriminant method, which shows better face recognition effects than some commonly used linear discriminant algorithms. The radial basis function (RBF) neural network is widely applied to the function approximation and pattern classification. One of the interesting research topics of RBF network is how to set appropriate hidden-layer units. Based on DCV, we design a new nonlinear feature extraction algorithm that is the kernel DCV (KDCV) algorithm and we employ the DCV generated by KDCV as the hidden-layer units of the RBF network. Then we present a novel face recognition approach that is the KDCV-RBF approach. Testing on a public large face database (AR database), the experimental results demonstrate that KDCV-RBF is an effective face recognition approach, which outperforms several representative recognition methods.

Using backpropagation neural network for face recognition with 2D+3D hybrid information

Biometric measurements received an increasing interest for security applications in the last two decades. After the 911 terrorist attacks, face recognition has been an active research in this area. However, very few research group focus on face recognition from both 2D and 3D facial images. Almost all existing recognition systems rely on a single type of face information: 2D intensity (or color) image or 3D range data set [Wang, Y., Chua, C., & Ho, Y. (2002). Facial feature detection and face recognition from 3D and 3D images. Pattern Recognition Letters, 23, 1191-1202]. The objective of this study is to develop an effective face recognition system that extracts and combines 2D and 3D face features to improve the recognition performance. The proposed method derived the information of 3D face (disparity face) using a designed synchronous Hopfield neural network. Then, we retrieved 2D and 3D face features with principle component analysis (PCA) and local autocorrelation coefficient (LAC) respectively. Eventually, the information of features was learned and classified using backpropagation neural networks. An experiment was conducted with 100 subjects, and for each subject thirteen stereo face images were taken with different expressions. Among them, seven faces with expressions were used for training, and the rest of the expressions were used for testing. The experimental results show that the proposed method effectively improved the recognition rate by combining the 2D with 3D face information.

Linear local tangent space alignment and application to face recognition

In this paper, linear local tangent space alignment (LLTSA), as a novel linear dimensionality reduction algorithm, is proposed. It uses the tangent space in the neighborhood of a data point to represent the local geometry, and then aligns those local tangent spaces in the low-dimensional space which is linearly mapped from the raw high-dimensional space. Since images of faces often belong to a manifold of intrinsically low dimension, we develop LLTSA algorithm for effective face manifold learning and recognition. Comprehensive comparisons and extensive experiments show that LLTSA achieves much higher recognition rates than a few competing methods.