Searcharxiv⌕ Search

arXiv subjects

Dong Chen

Publications and source records attributed to Dong Chen.

At least 289 records · Page 16Linked to original sources
arXiv2018

Exploring Hypergraph Representation on Face Anti-spoofing Beyond 2D Attacks

Face anti-spoofing plays a crucial role in protecting face recognition systems from various attacks. Previous model-based and deep learning approaches achieve satisfactory performance for 2D face spoofs, but remain limited for more advanced 3D attacks such as vivid masks. In this paper, we address 3D face anti-spoofing via the proposed Hypergraph Convolutional Neural Networks (HGCNN). Firstly, we construct a computation-efficient and posture-invariant face representation with only a few key points on hypergraphs. The hypergraph representation is then fed into the designed HGCNN with hypergraph convolution for feature extraction, while the depth auxiliary is also exploited for 3D mask anti-spoofing. Further, we build a 3D face attack database with color, depth and infrared light information to overcome the deficiency of 3D face anti-spoofing data. Experiments show that our method achieves the state-of-the-art performance over widely used 3D and 2D databases as well as the proposed one under various tests.

cs.CV↗
arXiv2018

A novel active learning framework for classification: using weighted rank aggregation to achieve multiple query criteria

Multiple query criteria active learning (MQCAL) methods have a higher potential performance than conventional active learning methods in which only one criterion is deployed for sample selection. A central issue related to MQCAL methods concerns the development of an integration criteria strategy (ICS) that makes full use of all criteria. The conventional ICS adopted in relevant research all facilitate the desired effects, but several limitations still must be addressed. For instance, some of the strategies are not sufficiently scalable during the design process, and the number and type of criteria involved are dictated. Thus, it is challenging for the user to integrate other criteria into the original process unless modifications are made to the algorithm. Other strategies are too dependent on empirical parameters, which can only be acquired by experience or cross-validation and thus lack generality; additionally, these strategies are counter to the intention of active learning, as samples need to be labeled in the validation set before the active learning process can begin. To address these limitations, we propose a novel MQCAL method for classification tasks that employs a third strategy via weighted rank aggregation. The proposed method serves as a heuristic means to select high-value samples of high scalability and generality and is implemented through a three-step process: (1) the transformation of the sample selection to sample ranking and scoring, (2) the computation of the self-adaptive weights of each criterion, and (3) the weighted aggregation of each sample rank list. Ultimately, the sample at the top of the aggregated ranking list is the most comprehensively valuable and must be labeled. Several experiments generating 257 wins, 194 ties and 49 losses against other state-of-the-art MQCALs are conducted to verify that the proposed method can achieve superior results.

cs.LG↗
arXiv2018

Towards Open-Set Identity Preserving Face Synthesis

We propose a framework based on Generative Adversarial Networks to disentangle the identity and attributes of faces, such that we can conveniently recombine different identities and attributes for identity preserving face synthesis in open domains. Previous identity preserving face synthesis processes are largely confined to synthesizing faces with known identities that are already in the training dataset. To synthesize a face with identity outside the training dataset, our framework requires one input image of that subject to produce an identity vector, and any other input face image to extract an attribute vector capturing, e.g., pose, emotion, illumination, and even the background. We then recombine the identity vector and the attribute vector to synthesize a new face of the subject with the extracted attribute. Our proposed framework does not need to annotate the attributes of faces in any way. It is trained with an asymmetric loss function to better preserve the identity and stabilize the training process. It can also effectively leverage large amounts of unlabeled training face images to further improve the fidelity of the synthesized faces for subjects that are not presented in the labeled training face dataset. Our experiments demonstrate the efficacy of the proposed framework. We also present its usage in a much broader set of applications including face frontalization, face attribute morphing, and face adversarial example detection.

cs.CV↗
arXiv2018

Unique equilibrium states for geodesic flows over surfaces without focal points

In this paper, we study dynamics of geodesic flows over closed surfaces of genus greater than or equal to 2 without focal points. Especially, we prove that there is a large class of potentials having unique equilibrium states, including scalar multiples of the geometric potential, provided the scalar is less than 1. Moreover, we discuss ergodic properties of these unique equilibrium states. We show these unique equilibrium states are Bernoulli, and weighted regular periodic orbits are equidistributed relative to these unique equilibrium states.

math.DS↗
arXiv2018

Weakly nonlinear analysis for car-following model with consideration of cooperation and time delays

In traffic systems, cooperative driving has attracted the researchers attentions. A lot of works attempt to understand the effects of cooperative driving behavior and/or time delays on traffic flow dynamics for specific traffic flow model. This paper is a new attempt to investigate analyses of linear stability and weak nonlinear for the general car-following model with consideration of cooperation and time delays. We derive linear stability condition and study that how the combinations of cooperation and time delays affect the stability of traffic flow. Burgers equation and Korteweg de Vries (KdV) equation for car-following model considering cooperation and time delays are derived. Their solitary wave solutions and constraint conditions are concluded. We investigate the property of cooperative optimal velocity(OV) model which estimates the combinations of cooperation and time delays about the evolution of traffic waves using both analytic and numerical methods. The results indicate that delays and cooperation are model-dependent, and cooperative behavior could inhibit the stabilization of traffic flow. Moreover, delays of sensing to relative motion are easy to trigger the traffic waves; delays of sensing to host vehicle are beneficial to relieve the instability effect a certain extent.

nlin.PS↗
arXiv2017

CVAE-GAN: Fine-Grained Image Generation through Asymmetric Training

We present variational generative adversarial networks, a general learning framework that combines a variational auto-encoder with a generative adversarial network, for synthesizing images in fine-grained categories, such as faces of a specific person or objects in a category. Our approach models an image as a composition of label and latent attributes in a probabilistic model. By varying the fine-grained category label fed into the resulting generative model, we can generate images in a specific category with randomly drawn values on a latent attribute vector. Our approach has two novel aspects. First, we adopt a cross entropy loss for the discriminative and classifier network, but a mean discrepancy objective for the generative network. This kind of asymmetric loss function makes the GAN training more stable. Second, we adopt an encoder network to learn the relationship between the latent space and the real image space, and use pairwise feature matching to keep the structure of generated images. We experiment with natural images of faces, flowers, and birds, and demonstrate that the proposed models are capable of generating realistic and diverse samples with fine-grained category labels. We further show that our models can be applied to other tasks, such as image inpainting, super-resolution, and data augmentation for training better face recognition models.

cs.CV↗
arXiv2017

Neural Aggregation Network for Video Face Recognition

This paper presents a Neural Aggregation Network (NAN) for video face recognition. The network takes a face video or face image set of a person with a variable number of face images as its input, and produces a compact, fixed-dimension feature representation for recognition. The whole network is composed of two modules. The feature embedding module is a deep Convolutional Neural Network (CNN) which maps each face image to a feature vector. The aggregation module consists of two attention blocks which adaptively aggregate the feature vectors to form a single feature inside the convex hull spanned by them. Due to the attention mechanism, the aggregation is invariant to the image order. Our NAN is trained with a standard classification or verification loss without any extra supervision signal, and we found that it automatically learns to advocate high-quality face images while repelling low-quality ones such as blurred, occluded and improperly exposed faces. The experiments on IJB-A, YouTube Face, Celebrity-1000 video face recognition benchmarks show that it consistently outperforms naive aggregation methods and achieves the state-of-the-art accuracy.

cs.CV↗
arXiv2017

Double-Helical Tiled Chain Structure of the Twist-Bend Liquid Crystal phase in CB7CB

The twist-bend nematic liquid crystal phase is a three-dimensional fluid in which achiral bent molecules spontaneously form an orientationally ordered macroscopically chiral heliconical winding of molecular scale pitch, in absence of positional ordering. Here we characterize the structure of the ground state of the twist-bend phase of the bent dimer CB7CB and its mixtures with 5CB over a wide range of concentrations and temperatures, showing that the contour length along the molecular direction for a single turn of the helix is approximately equal to 2πRmol, where Rmol is the radius of bend curvature of a single all-trans CB7CB molecule. This relation emerges from a model which simply relates the macroscopic characteristics of the helical structure, which is mostly biaxial twist and has little bend, to the bent molecular shape. This connection comes about through the presence in the fluid of self-assembled oligomer-like correlations of interlocking molecules, arising from the nanosegregation of rigid and flexible molecular subcomponents, forming a brickwork tiling of pairs of molecular strands into a duplex double-helical chain.

cond-mat.soft↗
arXiv2016

Positive metric entropy arises in some nondegenerate nearly integrable systems

The celebrated KAM Theory says that if one makes a small perturbation of a non-degenerate completely integrable system, we still see a huge measure of invariant tori with quasi-periodic dynamics in the perturbed system. These invariant tori are known as KAM tori. What happens outside KAM tori draws a lot of attention. In this paper we present a Lagrangian perturbation of the geodesic flow on a flat 3-torus. The perturbation is $C^\infty$ small but the flow has a positive measure of trajectories with positive Lyapunov exponent, namely, the flow has positive metric entropy. From this result we get positive metric entropy outside some KAM tori.

math.DS↗
arXiv2016

Supervised Transformer Network for Efficient Face Detection

Large pose variations remain to be a challenge that confronts real-word face detection. We propose a new cascaded Convolutional Neural Network, dubbed the name Supervised Transformer Network, to address this challenge. The first stage is a multi-task Region Proposal Network (RPN), which simultaneously predicts candidate face regions along with associated facial landmarks. The candidate regions are then warped by mapping the detected facial landmarks to their canonical positions to better normalize the face patterns. The second stage, which is a RCNN, then verifies if the warped candidate regions are valid faces or not. We conduct end-to-end learning of the cascaded network, including optimizing the canonical positions of the facial landmarks. This supervised learning of the transformations automatically selects the best scale to differentiate face/non-face patterns. By combining feature maps from both stages of the network, we achieve state-of-the-art detection accuracies on several public benchmarks. For real-time performance, we run the cascaded network only on regions of interests produced from a boosting cascade face detector. Our detector runs at 30 FPS on a single CPU core for a VGA-resolution image.

cs.CV↗
arXiv2015

The twist-bend nematic phase of bent mesogenic dimer CB7CB and its mixtures

Binary mixtures of the twist-bend nematic-forming liquid crystal CB7CB with the prototypical rod-like liquid crystal 5CB exhibit a twist-bend nematic phase with properties similar to those reported for neat CB7CB. The mixtures appear homogeneous, with no micron- or nano-scale segregation evident at any concentration. The linear dependence of the phase transition temperature on concentration indicates that these binary mixtures are nearly ideal. However, a decrease in the viscosity with the addition of 5CB allows the characteristic twist-bend stripe textures to relax into a state of uniform birefringence. We confirm the presence of nanoscale modulations of the molecular orientation in the mixtures by freeze-fracture transmission electron microscopy (FFTEM), further evidence of their twist-bend nature. We devise and implement a statistical approach to quantitatively measure the ground state pitch of the twist-bend phase and its mixtures using FFTEM. The addition of 5CB generally shifts the measured ground-state pitch distributions towards larger pitch. Interestingly, the pitch appears to increase discontinuously by ~10 nm at the 50 wt% concentration of 5CB, indicating that the twist-bend phase undergoes a structural transition at higher 5CB concentrations.

cond-mat.soft↗
arXiv2015

A characterization of L(2, 1)-labeling number for trees with maximum degree 3

An L(2, 1)-labeling of a graph is an assignment of nonnegative integers to the vertices of G such that adjacent vertices receive numbers differed by at least 2, and vertices at distance 2 are assigned distinct numbers. The L(2, 1)-labeling number is the minimum range of labels over all such labeling. It was shown by Griggs and Yeh [Labelling graphs with a condition at distance 2, SIAM J. Discrete Math. 5(1992), 586-595] that the L(2, 1)-labeling number of a tree is either \D+ 1 or \D + 2. In this paper, we give a complete characterization of L(2, 1)-labeling number for trees with maximum degree 3.

math.CO↗
arXiv2015

Observation of the chiral anomaly induced negative magneto-resistance in 3D Weyl semi-metal TaAs

Weyl semi-metal is the three dimensional analog of graphene. According to the quantum field theory, the appearance of Weyl points near the Fermi level will cause novel transport phenomena related to chiral anomaly. In the present paper, we report the first experimental evidence for the long-anticipated negative magneto-resistance generated by the chiral anomaly in a newly predicted time-reversal invariant Weyl semi-metal material TaAs. Clear Shubnikov de Haas oscillations (SdH) have been detected starting from very weak magnetic field. Analysis of the SdH peaks gives the Berry phase accumulated along the cyclotron orbits to be π, indicating the existence of Weyl points.

cond-mat.mtrl-sci↗
arXiv2013

Twist-bend heliconical chiral nematic liquid crystal phase of an achiral rigid bent-core mesogen

The chiral, heliconical (twist-bend) nematic ground state is reported in an achiral, rigid, bent-core mesogen (UD68). Similar to the nematic twist-bend (NTB) phase observed in bent molecular dimers, the NTB phase of UD68 forms macroscopic, smectic-like focal-conic textures and exhibits nanoscale, periodic modulation with no associated modulation of the electron density, i.e., without a detectable lamellar x-ray reflection peak. The NTB helical pitch is pTB ~ 14 nm. When an electric field is applied normal to the helix axis, a weak electroclinic effect is observed, revealing 50 um-scale left- and right-handed domains in a chiral conglomerate.

cond-mat.mtrl-sci↗
arXiv2013

A Twist-Bend Chiral Helix of 8nm Pitch in a Nematic Liquid Crystal of Achiral Molecular Dimers

Freeze Fracture Transmission Electron Microscopy (FFTEM) study of the nanoscale structure of the so-called "twist-bend" nematic (NX) phase of the cyanobiphenyl (CB) dimer molecule CB(CH2)7CB reveals a stripe texture of fluid layers periodically arrayed with a bulk spacing of d ~ 8.3 nm. Fluidity and a rigorously maintained spacing produce long-range-ordered fluid layered focal conic domains. Absence of a lamellar x-ray reflection at wavevector q ~ 2π/8 nm-1 or its harmonics in synchrotron-based scattering experiments indicates that this periodic structure is achieved with no detectable associated modulation of the electron density, and thus has nematic molecular ordering. A search for periodic ordering with d ~ 8nm in CB(CH2)7CB using atomistic molecular dynamic computer simulation yielded equilibration of a conical twist-bend helixed nematic ground state, of the sort first proposed by Meyer, and envisioned in systems of bent molecules by Dozov and Memmer, We identify θ ~ 33 degree as the cone angle, and p ~ 8nm as the full pitch of the helix, the shortest ever found in a nematic fluid.

cond-mat.soft↗
arXiv2012

Single and multiple index functional regression models with nonparametric link

Fully nonparametric methods for regression from functional data have poor accuracy from a statistical viewpoint, reflecting the fact that their convergence rates are slower than nonparametric rates for the estimation of high-dimensional functions. This difficulty has led to an emphasis on the so-called functional linear model, which is much more flexible than common linear models in finite dimension, but nevertheless imposes structural constraints on the relationship between predictors and responses. Recent advances have extended the linear approach by using it in conjunction with link functions, and by considering multiple indices, but the flexibility of this technique is still limited. For example, the link may be modeled parametrically or on a grid only, or may be constrained by an assumption such as monotonicity; multiple indices have been modeled by making finite-dimensional assumptions. In this paper we introduce a new technique for estimating the link function nonparametrically, and we suggest an approach to multi-index modeling using adaptively defined linear projections of functional data. We show that our methods enable prediction with polynomial convergence rates. The finite sample performance of our methods is studied in simulations, and is illustrated by an application to a functional regression problem.

math.ST↗
arXiv2012

Nonlinear manifold representations for functional data

For functional data lying on an unknown nonlinear low-dimensional space, we study manifold learning and introduce the notions of manifold mean, manifold modes of functional variation and of functional manifold components. These constitute nonlinear representations of functional data that complement classical linear representations such as eigenfunctions and functional principal components. Our manifold learning procedures borrow ideas from existing nonlinear dimension reduction methods, which we modify to address functional data settings. In simulations and applications, we study examples of functional data which lie on a manifold and validate the superior behavior of manifold mean and functional manifold components over traditional cross-sectional mean and functional principal components. We also include consistency proofs for our estimators under certain assumptions.

math.ST↗