Abstract — We consider the problem of automatically recognizing human faces from frontal views with varying expression and illumination, as well as occlusion and disguise. We cast the recognition problem as one of classifying among multiple linear regression models, and argue that new theory from sparse signal representation offers the key to addressing this problem. Based on a sparse representation computed by ℓ 1-minimization, we propose a general classification algorithm for (image-based) object recognition. This new framework provides new insights into two crucial issues in face recognition: feature extraction and robustness to occlusion. For feature extraction, we show that if sparsity in the recognition problem is properly harnessed, the choice of features is no longer critical. What is critical, however, is whether the number of features is sufficiently large and whether the sparse representation is correctly computed. Unconventional features such as downsampled images and random projections perform just as well as conventional features such as Eigenfaces and Laplacianfaces, as long as the dimension of the feature space surpasses certain threshold, predicted by the theory of sparse representation. This framework can handle errors due to occlusion and corruption uniformly, by exploiting the fact that these errors are often sparse w.r.t. to the standard (pixel) basis. The theory of sparse representation helps predict how much occlusion the recognition algorithm can handle and how to choose the training images to maximize robustness to occlusion. We conduct extensive experiments on publicly available databases to verify the efficacy of the proposed algorithm, and corroborate the above claims.

Face recognition is one of the most intensively studied topics in computer vision and pattern recognition. Facial expression, which changes face geometry, usually has an adverse effect on the performance of a face recognition system. On the other hand, face geometry is a useful cue for recognition. Taking these into account, we utilize the idea of separating geometry and texture information in a face image and model the two types of information by projecting them into separate PCA spaces which are specially designed to capture the distinctive features among different individuals. Subsequently, the texture and geometry attributes are re-combined to form a classifier which is capable of recognizing faces with different expressions. Finally, by studying face geometry, we are able to determine which type of facial expression has been carried out, thus build an expression classifier. Numerical validations of the proposed method are given. 1.

Abstract. This paper investigates the main reason for the obtained low performance when the face recognition algorithms are tested on partially occluded face images. It has been observed that in the case of upper face occlusion, missing discriminative information due to occlusion only accounts for a very small part of the performance drop. The main factor is found to be the registration errors due to erroneous facial feature localization. It has been shown that by solving the misalignment problem, very high correct recognition rates can be achieved with a generic local appearance-based face recognition algorithm. In the case of a lower face occlusion, only a slight decrease in the performance is observed, when a local appearance-based face representation approach is used. This indicates the importance of local processing when dealing with partial face occlusion. Moreover, improved alignment increases the correct recognition rate also in the experiments against the lower face occlusion, which shows that face registration plays a key role on face recognition performance. 1

We investigate the use of multiple intrinsic geometric attributes, including angles, geodesic distances, and curvatures, for 3D face recognition, where each face is represented by a triangle mesh, preprocessed to possess a uniform connectivity. As invariance to facial expressions holds the key to improving recognition performance, we propose to train for the component-wise weights to be applied to each individual attribute, as well as the weights used to combine the attributes, in order to adapt to expression variations. Using the eigenface approach based on the training results and a nearest neighbor classifier, we report recognition results on the expression-rich GavabDB face database and the well-known Notre Dame FRGC 3D database. We also perform a cross validation between the two databases.

Subspace projection techniques are known to be susceptible to the presence of partial occlusions in the image data. To overcome this susceptibility, we present in this paper a confidence weighting scheme that assigns weights to pixels according to a measure, which quantifies the confidence that the pixel in question represents an outlier. With this procedure the impact of the occluded pixels on the subspace representation is reduced and robustness to partial occlusions is obtained. Next, the confidence weighting concept is improved by a local procedure for the estimation of the subspace representation. Both the global weighting approach and the local estimation procedure are assessed in face recognition experiments on the AR database, where encouraging results are obtained with partially occluded facial images. 1.

Abstract — In this paper, we present a face registration approach, in which alignment is done by minimizing the closest distance at the classification step. This method eliminates the need of a feature localization step that exists in traditional face recognition systems and formulates alignment as an optimization process during classification. In other words, instead of performing a separate facial feature localization step and localizing facial features according to some type of feature matching score, in the proposed method, alignment is done by directly optimizing the classification score. Moreover, a feature detector can still be integrated to the system. In this case, the output of the feature detector is used as the initial point of the optimization process. Results of extensive experiments have shown that the proposed approach leads very high correct recognition rates, especially in the case of partial face occlusion, where it is not possible to precisely detect the facial feature locations. It has been also found that, in the case of using a facial feature detector, the approach can tolerate localization errors of up to 18 % of the interocular distance. I.

This paper develops a novel face recognition technique