Abstract—A number of current face recognition algorithms use face representations found by unsupervised statistical methods. Typically these methods find a set of basis images and represent faces as a linear combination of those images. Principal component analysis (PCA) is a popular example of such methods. The basis images found by PCA depend only on pairwise relationships between pixels in the image database. In a task such as face recognition, in which important information may be contained in the high-order relationships among pixels, it seems reasonable to expect that better basis images may be found by methods sensitive to these high-order statistics. Independent component analysis (ICA), a generalization of PCA, is one such method. We used a version of ICA derived from the principle of optimal information transfer through sigmoidal neurons. ICA was performed on face images in the FERET database under two different architectures, one which treated the images as random variables and the pixels as outcomes, and a second which treated the pixels as random variables and the images as outcomes. The first architecture found spatially local basis images for the faces. The second architecture produced a factorial face code. Both ICA representations were superior to representations based on PCA for recognizing faces across days and changes in expression. A classifier that combined the two ICA representations gave the best performance. Index Terms—Eigenfaces, face recognition, independent component analysis (ICA), principal component analysis (PCA), unsupervised learning. I.

This article addresses the issue of expressive face modelling using an active appearance model for facial expression recognition and synthesis. We consider the six universal emotional categories namely joy, anger, fear, disgust, sadness and surprise. After a description of the active appearance model (computed with 3 or only one PCA), we address the active appearance model contribution to automatic facial expression recognition. Then we propose a new method for analysis and synthesis allowing, from a single photo, to cancel the facial expression on a given face and to artificially synthesize novel expressions on this same face. In this last framework, we propose two facial expression modelling approaches.

A number of current face recognition algorithms use face representations found by unsupervised statistical methods. Typically these methods find a set of basis images and represent faces as a linear combination of those images. Principal component analysis (PCA) is a popular example of such methods. The basis images found by PCA depend only on pair-wise relationships between pixels in the image database. In a task such as face recognition, in which important information may be contained in the high-order relationships among pixels, it seems reasonable to expect that better basis images may be found by methods sensitive to these high order statistics. Independent component analysis (ICA), a generalization of PCA, is one such method. We used a version of ICA derived from the principle of maximum information transfer through sigmoidal neurons [12]. ICA was performed on face images in the FERET database under two different architectures, one which treated the images as random variables and the pixels as outcomes, and a second which treated the pixels as random variables and the images as outcomes. The first architecture found spatially local basis images for the faces. The second architecture produced a factorial face code. Both ICA representations were superior to representations based on principal component analysis for recognizing faces across

In this chapter we present a system for face recognition that combines two recent advances in computer graphics and computer vision: 3D morphable models and component-based recognition. By fitting a morphable model to a triplet of face images we generate a 3D head model for each person in our face database. The 3D models are rendered under varying pose and illumination conditions to build a large set of synthetic images. We then train a component-based face recognition system on these synthetic images. At runtime, the face recognition module is preceded by a hierarchical face detector resulting in a system that can detect and identify faces in video images at about 4 Hz. The system achieved a recognition rate which was significantly higher than that of a comparable global face recognition system trained on the same data. Finally, we address the problem of how to automatically determine the size and shape of facial components for face identification. The need for a robust, accurate, and easily trainable face recognition system becomes more pressing as real world applications in the areas of law enforcement, surveillance, access control, and human machine interfaces continue to develop. However, extrinsic imaging parameters such as pose,

Abstract—When the feature dimension is larger than the number of samples the small sample-size problem occurs. There is great concern about it within the face recognition community. We point out that optimizing the Fisher index in linear discriminant analysis does not necessarily give the best performance for a face recognition system. We propose a new regularization scheme. The proposed method is evaluated using the Olivetti Research Laboratory database, the Yale database, and the Feret database. Index Terms—Face recognition, optimization, regularized discriminant analysis (RDA), small sample-size problem. I.

In this paper, we present a novel feature extraction called the protoface. While the Eigenface is based on principal component analysis and the Fisherface on Fisher's linear discriminant analysis, the protoface only requires decomposing the covariance matrix of the prototypes that can better describe whole observations. It is thus more computationally e#cient especially when the number of the observations is large. Moreover, instead of using the method of multivariate analysis of variance, protoface relies on the analysis of distance. It is thus more appropriate in situations where the assumptions of normality of the observations are violated. Besides, by using the kernel trick, we also obtain a kernel protoface in the feature space defined by the kernel. Finally, we apply our methods to face recognition, and experimental results on both the AT&T and Yale databases are reported.

This paper presents how to extract non-linear features by linear PCA. KPCA is effective but the computational cost is the drawback. To realize both non-linearity and low computational cost simultaneously, the idea of local kernel is used. The mapped features of the polynomial kernel can be described explicitly. When input features are divided into some local features and the polynomial kernel is applied to each local features independently, the dimension of mapped features does not become so high. In addition, the inner product with all local mapped features corresponds to the local summation kernel. Thus, KPCA with the local summation kernel can be solved by linear PCA. The proposed approach is evaluated in object categorization problem which requires high non-linearity and computational cost. The proposed method gives much higher accuracy than linear PCA. The computational cost is lower than KPCA though the accuracy is slightly worse than KPCA. 1.

A key issue that continues to generate controversy concerns the nature of the psychological, computational, and neural mechanisms that support the visual recognition of objects such as faces and words. While some researchers claim that visual recognition is accomplished by category-specific modules dedicated to processing distinct object classes, other researchers have argued for a more distributed system with only partially specialized cortical regions. Considerable evidence from both functional neuroimaging and neuropsychology would seem to favour the modular view, and yet close examination of those data reveals rather graded patterns of specialization that support a more distributed account. This paper explores a theoretical middle ground in which the functional specialization of brain regions arises from general principles and constraints on neural representation and learning that operate throughout cortex but that nonetheless have distinct implications for different classes of stimuli. The account is supported by a computational simulation, in the form of an artificial neural network, that illustrates how cooperative and competitive interactions in the formation of neural representations for faces and words account for both their shared and distinctive properties. We set out a series of empirical predictions, which are also examined, and consider the further implications of this account.