Partial Face Recognition Based on Template Matching

Component-based face recognition under transfer learning for forensic applications

A single gallery-based face recognition using extended joint sparse representation

Matching partial heterogeneous face images to a gallery of visible images is a challenging research problem. This scenario is motivated by a number of surveillance applications such as recognition of subjects at night or in the presence of challenging environmental conditions. Standoff distances may range from a meter to hundred meters. Our latest experiments have shown that a face recognition software recently developed in our research group can be adapted to perform cross spectral matching of partial face images. The images are encoded with Gabor Generalized Local Binary Patterns and Gabor Weber operators and matched by means of a Kullbuck-Leibler metric. Our analysis has shown that three separate face regions such as (1) eyes and nasal bridge, (2) cheeks and nasal tip, and (3) mouth and a part of the chin display similar matching performance. Furthermore, we have evaluated performance of periocular regions. For a short standoff distance of 1.5 meters and a database of 48 classes, matching a Short Wave Infrared (SWIR) periocular region against visible regions resulted in 0.7 Genuine Accept Rate (GAR) at False Accept Rate (FAR) set to 0.01. For a long standoff distance of 106 meters and a database of 48 classes, matching SWIR against visible periocular regions yielded 0.4 GAR at FAR equal to 0.1. For a short standoff distance of 1.5 meters and a database of 200 classes, matching a Medium Wave Infrared (MWIR) periocular region against visible regions resulted in 0.35 GAR at FAR set to 0.1.

Gender classification from facial images plays a significant role in biometric technology viz. gender medicine, surveillance, electronic banking system and human computer interaction. However, it has many challenges due to variations of pose, expression, aging, race, make-up, occlusion and illumination. In the proposed system, spatially enhanced local binary pattern (SLBP) and histogram of oriented gradients (HOG) are extracted to classify the human gender with SVM classifier. This hybrid feature selection has increased the power of the proposed system due to its representation of texture micro-patterns and local shape by capturing the edge or gradient structure form the image. The gender classification accuracy is studied by using the local feature representation of the face images separately and also these features are concatenated to provide a better recognition rate. The combination of two different local descriptors provides good representation of face image and this is given to SVM classifier which classifies as male or female. Also, the proposed work is compared with other two traditional classifiers such as k-nearest neighbor and sparse representation classifier. The performance was evaluated on FERET and LFW database. The highest classification accuracy 99.1% is achieved on FERET database and 95.7% is achieved on LFW database by applying cubic SVM with fusion of SLBP and HOG features.

The purpose of this research is to present Symmetrical Singular Value Decomposition Representation (SSVDR) method with Gabor Filter under illumination effect for face recognition. SSVDR method was proposed to normalize illuminated face images caused by the difference of light. SSVDR method represented characteristics of face images on low-intensity combined with the symmetrical reversed part of the face based on Singular Value Decomposition (SVD). Gabor Filter was used to extract the face images which were already processed using SSVDR. In order to do the face recognition, PCA and LDA methods were used with Nearest Neighbor as the classifier. The result showed that face recognition with SSVDR based on Gabor Filter was quite good on face images with low-intensity, but the recognition were not good enough on face images with extreme illumination. Overall, the accuracy of face recognition scored 91.86% with PCA and 91.57% with LDA.

Variations in lighting conditions, pose and expression make face recognition an even more challenging and difficult task. This paper presents a face recognition approach by using image enhancement and Gabor wavelets transformation. In face recognition, image preprocessing is the key step since it is important to features extract and recognition. Logarithm transformation and normalization are performed in face images captured under various lighting conditions for face recognition. This involves convolving a face image with a series of Gabor wavelets at different scales, locations, and orientations and extracting features from Gabor filtered images. Significant improvements are also observed when the preprocessing and Gabor filtered images are used for feature extraction instead of the original images. The approach achieves 94.4% recognition accuracy using only 160 features of a face image. Experimental results show that the proposed approach improves face recognition performance using this scheme when training and testing on images captured under variable illumination and expression.

Nowadays, computer-based face recognition is a mature and reliable mechanism that is significantly used in many access control scenarios along with other biometric methods. Face recognition consists of two subtasks including Face Verification and Face Identification. By comparing a pair of images, Face Verification determines whether those images are related to one person or not; and Face Identification has to identify a specific face within a set of available faces in the database. There are many challenges in face recognition such as angle, illumination, pose, facial expression, noise, resolution, occlusion and the few number of one-class samples with several classes. In this paper, we are carrying out face recognition by utilizing transfer learning in a siamese network which consists of two similar CNNs. In the siamese network, a pair of two face images is given to the network as input, then the network extracts the features of this pair of images and finally, it determines whether the pair of images belongs to one person or not by using a similarity criterion. The results show that the proposed model is comparable with advanced models that are trained on datasets containing large numbers of samples. furthermore, it improves the accuracy of face recognition in comparison with methods which are trained using datasets with a few number of samples, and the mentioned accuracy is claimed to be 95.62% on LFW dataset.

Real-time face detection and recognition using hybrid-information extracted from face space and facial features

A comprehensive account of almost all the feature extraction methods used in current face recognition systems is presented. Specifically we have made distinction in the holistic and local feature extraction and differentiate them qualitatively as opposed to quantitatively. It is argued that a global feature representation should be preferred over a bag-of-feature approach. The problems in current feature extraction techniques and their reliance on a strict alignment is discussed. Finally we have introduced to use face-GLOH signatures that are invariant with respect to scale, translation and rotation and therefore do not require properly aligned images. The resulting dimensionality of the vector is also low as compared to other commonly used local features such as Gabor, LBP etc. and therefore learning based methods can also benefit from it. In a typical multi-view face recognition task, where it is assumed to have several examples of a subject available for training, we have shown in an extensive experimental setting the advantages and weaknesses of commonly used feature descriptions. Our results show that under more realistic assumptions, most of the classifiers failed on conventional features. While using the introduced face-GLOH-signature representation is relatively less affected by large in-class variations. This has been demonstrated by providing a fair performance comparison of several classifiers under more practical conditions such as misalignments, large number of subjects and large pose variations. An important conclusion is to be drawn from the results on FERET is that conventional multi-view face recognition cannot cope well with regards to large pose variations. Even using a large number of training examples in different poses for a subject do not suffice for a satisfactory recognition. In order to solve the problem where only one training example per subject is available, many recent methods propose to use image synthesis to generate a given subject at all other views and then perform a conventional multi-view recognition (Beymer and Poggio, 1995; Gross et al, 2004). Besides the fact that such synthesis techniques cause severe artefacts and thus cannot preserve the identity of an individual, a conventional classification cannot yield good recognition results, as has been shown in an extensive experimental setting. More sophisticated methods are therefore needed in order to address pose invariant face recognition. Large pose differences cause significant appearance variations that in general are larger than the appearance variation due to identity. One possible way of addressing this is to learn these variations across each pose, more specifically by fixing the pose and establishing a correspondence on how a person’s appearance changes under this pose one could reduce the in-class appearance variation significantly. In our very recent work (Sarfraz and Hellwich, 2009), we demonstrate the usefulness of face-GLOH signature in this direction.

Face alignment and face recognition are two fundamental problems in the facial analysis community. For face alignment, it forms the basis for the accurate face recognition, age estimation, and facial expression recognition. For face recognition, it has been widely applied in various practical scenarios such as access control system, massive surveillance, human computer interaction, etc.. There mainly exist two lines of works in these two fields, namely holistic face alignment and recognition, and partial face alignment and recognition. Numerous holistic face alignment and recognition works

JPEG compression standard is widely used for reducing the volume of images that are stored or transmitted via networks. In biometrics datasets, facial images are usually stored in JPEG compressed format and should be fully decompressed to be used in a face recognition system. Recently, in order to reduce the computational complexity of JPEG decompression step, face recognition in compressed domain is considered as an emerging topic in face recognition systems. In this paper, a novel coefficient selection method based on face segmentation has been proposed for selecting a limited number of zigzag scanned quantized coefficients in JPEG compressed domain, which led to an improvement in recognition accuracy and a reduction in computational complexity of the face recognition system. In the proposed method, different low frequency coefficients based on the importance of the regions of a face have been selected for recognition process. The experiments were conducted on FERET and FEI face databases, and PCA and ICA methods have been utilized to extract the features of the selected coefficients. Different criteria including recognition accuracy and time complexity metrics were employed in order to evaluate the performance of the proposed method, and the results have been compared with those of the state-of-the art methods. The results show the superiority of the proposed approach, in terms of recognition ranks, discriminatory power and time complexity aspects.

A gradient-based face recognition method using Partial Hausdorff Distance (PHD) measure is proposed in this paper. First, in order to achieve a performance independent of lighting conditions, the image is transformed into a Gradient Map (GM). And then, Hausdorff distance measure is introduced to calculate the dissimilarity between two Gradient Maps. The experimental data show that the measure is suitable for face recognition. As we can see later, this distance measure is robust to lighting variations, slight pose differences and expression changes in face images. At last, recognition accuracy is given tested on AR and FERET databases, and comparisons with Edge Map (EM) and Line segment Edge Map (LEM) approaches are also presented.

Face images obtained by an outdoor surveillance camera, are often confronted with severe degradations (e.g., low-resolution, low-contrast, blur and noise). This significantly limits the performance of face recognition (FR) systems. This paper presents a framework to overcome the degradation in images obtained by an outdoor surveillance camera, to improve the performance of FR. We have defined a measure that is based on the difference in intensity histograms of face images, to estimate the amount of degradation. In the past, super-resolution techniques have been proposed to increase the image resolution for face recognition. In this work, we attempt a combination of partial restoration (using super-resolution, interpolation etc.) of probe samples (long distance shots of outdoor) and simulated degradation of gallery samples (indoor shots). Due to the unavailability of any benchmark face database with gallery and probe images, we have built our own database and conducted experiments on a realistic surveillance face database. PCA and FLDA have been used as baseline face recognition classifiers. The aim is to illustrate the effectiveness of our proposed method of compensating the degradation in surveillance data, rather than designing a specific classifier space suited for degraded test probes. The efficiency of the method is shown by improvement in the face classification accuracy, while comparing results obtained separately using training with acquired indoor gallery samples and then testing with the outdoor probes.KeywordsDegradationFace databaseFace recognitionImage qualitySurveillanceSuper-resolution

Face recognition is biometric pattern recognition, which is more acceptable and convenient for users compared with other biometric recognition traits. Among many problems in face recognition system, pose problem is still considered as one of the major problem still unsolved in satisfactory level. This paper proposes a novel pose tolerant face recognition approach. Proposed system's contribution is a novel and efficient approach that includes feature extraction, pose transformation learning and recognition stages. In first stage, 2D PCA is used as robust feature extraction technique. The linear regression is used as efficient and accurate transformation learning technique to create frontal face image from a different posed face images in the second stage. In last stage, Mahalanobis distance is used for recognition. Experiments on FERET and FEI face databases demonstrate higher performance in comparison with traditional systems.

A novel Gabor-Kernel face recognition method is proposed in this paper. This involves convolving a face image with a series of Gabor wavelets at different scales, locations, and orientations and extracting features from resulting Gabor filtered images. kernel discriminant analysis (KDDA) is then applied to the feature vectors for dimension reduction as well as class separability enhancement. A database of 600 frontal-view face images from the FERET face database is used to test the method. Experimental results demonstrate the advantage of KDDA over other Kernel methods such as kernel principal component analysis (KPCA) and general discriminant analysis (GDA). Significant improvements are also observed when features are extracted from Gabor filtered images instead of the original images. A 94% accuracy has been observed for the novel Gabor + KDDA method on the FERET database using a simple classifier, which could be further improved by employing a more complex classifier and distance measurer.