Search

Many recent variational optical flow methods are not robust for illumination variance, and they only consider local image relation in terms of illumination. In this paper, we propose a new efficient illumination-invariance total variation optical flow method called the weighted regularization transform, which uses and optimizes the Weber’s Law. Our method exploits unequal probability as the weight that has non-local information to estimate stable optical flow despite illumination changes. The proposed method uses a coarse-to-fine pyramid model to reduce the influence on the data term from illumination. Then, an energy optimization procedure is introduced to constrain the minimization of the data term with the non-local regularization. Experimentation with the proposed method has been performed on three optical flow datasets and a face liveness detection database, which have challenging illumination variations, and the results demonstrate that the proposed method is quite robust with respect to variations in illumination.

The single-image super-resolution techniques (SISR) have been significantly promoted by deep networks. However, the storage and computation complexities of deep models increase dramatically alongside with the reconstruction performance. This paper proposes a densely connected recursive network (DCRN) to trade off the performance and complexity. We introduce an enhanced dense unit by removing the batch normalization (BN) layers and employing the squeeze-and-excitation (SE) structure. A recursive architecture is also adopted to control the parameters of deep networks. Moreover, a de-convolution based residual learning method is proposed to accelerate the residual feature extraction process. The experimental results validate the efficiency of the proposed approach.

Age estimation based on unconstrained face images remains a challenging problem in computer vision and pattern recognition. We address this by ensembling part-based features and designing a cost sensitive loss to overcome the general imbalance data in age estimation. Specifically, we treat age estimation as a multi-class classification problem and mainly make two contributions: (i) We present a Part-based Convolutional Network(PCN) for age estimation to extract regional features instead of the holistic ones, which can preserve more discriminative features in facial regions like forehead, canthus, cheek, jaw, etc; (ii) A balanced loss for multi-class classification is designed to handle the extremely imbalance age distribution. The loss pays more attention to the hard examples, and can automatically adjust the weights of samples according to their contribution. State-of-the-art performance was achieved on FG-NET, MORPH and CACD databases, validating the effectiveness of the proposed approach.

It is generally known that the illumination could seriously affect the performance of face analysis algorithms. Moreover, in most practical applications, the illumination is usually uncontrolled. A number of methods have been put forward to tackle the problem of illumination variations in face images, but they always only work on facial region and need to segment faces in advance. Furthermore, many illumination processing methods only demonstrate on grayscale images and require strict alignment of face images, resulting in limited applications in the real world. In this paper, we propose a face image illumination processing method based on the Generative Adversarial Network (GAN) with dual triplet loss. Through considering the inter-domain similarity and intra-domain difference between the generated images and the real images, we put forward the dual triplet loss. At the same time, we introduce the self-similarity constraint of the images in the target illumination field. Experiments on the CMU Multi-PIE face datasets demonstrate that the proposed method preserve the facial details well when relighting. The experiment of 3D face reconstruction also verifies the effectiveness of the proposed method.

Visual versus near infrared (VIS-NIR) face image matching uses an NIR face image as the probe and conventional VIS face images as enrollment. It takes advantage of the NIR face technology in tackling illumination changes and low-light condition and can cater for more applications where the enrollment is done using VIS face images such as ID card photos. Existing VIS-NIR techniques assume that during classifier learning, the VIS images of each target people have their NIR counterparts. However, since corresponding VIS-NIR image pairs of the same people are not always available, which is often the case, so those methods cannot be applied. To address this problem, we propose a transductive method named transductive heterogeneous face matching (THFM) to adapt the VIS-NIR matching learned from training with available image pairs to all people in the target set. In addition, we propose a simple feature representation for effective VIS-NIR matching, which can be computed in three steps, namely Log-DoG filtering, local encoding, and uniform feature normalization, to reduce heterogeneities between VIS and NIR images. The transduction approach can reduce the domain difference due to heterogeneous data and learn the discriminative model for target people simultaneously. To the best of our knowledge, it is the first attempt to formulate the VIS-NIR matching using transduction to address the generalization problem for matching. Experimental results validate the effectiveness of our proposed method on the heterogeneous face biometric databases.

Face recognition is an increasingly popular technology for user authentication. However, face recognition is susceptible to spoofing attacks. Therefore, a reliable way to detect malicious attacks is crucial to the robustness of the face recognition system. This paper describes a new approach to utilizing light field camera for defending spoofing face attacks, like (warped) printed 2D facial photos and high-definition tablet images. The light field camera is a sensor that can record the directions as well as the colors of incident rays. Needing only one snapshot, multiple refocused images can be generated. In the proposed method, three kinds of features extracted from a pair of refocused images are extracted to discriminate fake faces and real faces. To verify the performance, we build a light field photograph databases and conduct experiments. Experimental results reveal that the employed features can achieve remarkable anti-spoofing accuracy under different types of spoofing attacks.

Person re-identification (re-id) aims at matching person images across multiple surveillance cameras. Currently, most re-id systems highly rely on color cues, which are only effective in good illumination conditions, but fail in low lighting conditions. However, for security issues, it is very important to conduct surveillance in low lighting conditions. To remedy this problem, we propose using depth cameras to perform surveillance in dark places, while using traditional RGB cameras in bright places. Such a heterogeneous camera network brings a challenge to match images across depth and RGB cameras. In this paper, we mine the correlation between two heterogeneous cues (depth and RGB) on both feature-level and transformation-level. As such, depth-based features and RGB-based features are transformed to the same space, which alleviates the problem of cross-modality matching between depth and RGB cameras. Experimental results on two benchmark heterogeneous person re-id datasets show the effectiveness of our method.

Partitioning nonlinearly separable datasets is a basic problem that is associated with data clustering. In this paper, a novel approach that is termed graph-based multiprototype competitive learning (GMPCL) is proposed to handle this problem. A graph-based method is employed to produce an initial, coarse clustering. After that, a multiprototype competitive learning is introduced to refine the coarse clustering and discover clusters of an arbitrary shape. The GMPCL algorithm is further extended to deal with high-dimensional data clustering, i.e., the fast graph-based multiprototype competitive learning (FGMPCL) algorithm. An experimental comparison has been performed by the exploitation of both synthetic and real-world datasets to validate the effectiveness of the proposed methods. Additionally, we apply our GMPCL/FGMPCL to two computer-vision tasks, namely, automatic color image segmentation and video clustering. Experimental results show that GMPCL/FGMPCL provide an effective and efficient tool with application to computer vision.

The affinity propagation (AP) clustering algorithm has received much attention in the past few years. AP is appealing because it is efficient, insensitive to initialization, and it produces clusters at a lower error rate than other exemplar-based methods. However, its single-exemplar model becomes inadequate when applied to model multisubclasses in some situations such as scene analysis and character recognition. To remedy this deficiency, we have extended the single-exemplar model to a multi-exemplar one to create a new multi-exemplar affinity propagation (MEAP) algorithm. This new model automatically determines the number of exemplars in each cluster associated with a super exemplar to approximate the subclasses in the category. Solving the model is NP--hard and we tackle it with the max-sum belief propagation to produce neighborhood maximum clusters, with no need to specify beforehand the number of clusters, multi-exemplars, and superexemplars. Also, utilizing the sparsity in the data, we are able to reduce substantially the computational time and storage. Experimental studies have shown MEAP's significant improvements over other algorithms on unsupervised image categorization and the clustering of handwritten digits.

The visual-near infrared (VIS-NIR) face matching, sharing the illumination-invariant property of NIR face image and remaining the use of existing VIS face images as enrollment, has been a popular issue in recent years. However, existing techniques assume that there are sufficient pairwise VIS and NIR images for each person during training, which is not realistic in VIS-NIR matching problem, as no NIR images are available for people who have already been registered in the existing face recognition system and only a handful of pairwise VIS and NIR face images captured from new people are available. To address this problem, we formulate the VIS-NIR matching as a transductive learning problem, which is a first attempt to our best knowledge. Moreover, we propose a transductive method named Transductive Heterogeneous Face Matching (THFM) by alleviating the domains difference and learning the discriminative model for target simultaneously, making it possible to take the query/probe NIR images into account in a transductive way. Experimental results validate the effectiveness of our approach on the heterogeneous face biometric database.

In the past decade, illumination problem has been the bottleneck of developing robust face recognition systems. Extracting illumination invariant features, especially the gradient based descriptor [13], is an effective tool to solve this issue. In this paper, we propose a novel gradient based descriptor, namely Complete Gradient Face (CGF), to compensate the limitations in [13] and contribute in three folds: (1) we incorporate homogeneous filtering to alleviate the illumination effect and enhance facial information based on the Lambertian assumption; (2) we demonstrate the gradient magnitude in logarithm domain is insensitive to lighting change; (3) we propose a histogram based feature descriptor to integrate both magnitude and orientation information. Experiments on CMU-PIE and Extended YaleB are conducted to verify the effectiveness of our proposed method.

As an emerging field of speech recognition, dialect identification plays an important role for promoting applications of speech recognition technology. Since the communications among Mainland China, Hong Kong and Taiwan are becoming frequently, it is particularly necessary to identify their dialects. This paper makes contributions to this issue in the following threefolds: 1) we build a speech corpus for main dialects of the three areas; 2) we use the popular GMM based method to extensively evaluate the main dialects between Mainland China and Hong Kong and the ones between Mainland China and Taiwan, and we find the differences between Mainland China Mandarin and Taiwan Mandarin are much smaller than those between Mandarin and Cantonese, resulting in unsatisfactory results in the latter case; 3) we propose an improved method based on the analysis of GMM, namely, maximum KL distance based Gaussian component selection (MKLD-GCS) in order to improve the performance of dialect identification between Mainland China Mandarin and Taiwan Mandarin. Experimental results show that our proposed method obtains better identification performance than related methods.