arXiv subjects

Qi Li

Publications and source records attributed to Qi Li.

At least 505 records · Page 28Linked to original sources

arXiv2018

Combining Data-driven and Model-driven Methods for Robust Facial Landmark Detection

Facial landmark detection is an important yet challenging task for real-world computer vision applications. This paper proposes an effective and robust approach for facial landmark detection by combining data- and model-driven methods. Firstly, a Fully Convolutional Network (FCN) is trained to compute response maps of all facial landmark points. Such a data-driven method could make full use of holistic information in a facial image for global estimation of facial landmarks. After that, the maximum points in the response maps are fitted with a pre-trained Point Distribution Model (PDM) to generate the initial facial shape. This model-driven method is able to correct the inaccurate locations of outliers by considering the shape prior information. Finally, a weighted version of Regularized Landmark Mean-Shift (RLMS) is employed to fine-tune the facial shape iteratively. This Estimation-Correction-Tuning process perfectly combines the advantages of the global robustness of data-driven method (FCN), outlier correction capability of model-driven method (PDM) and non-parametric optimization of RLMS. Results of extensive experiments demonstrate that our approach achieves state-of-the-art performances on challenging datasets including 300W, AFLW, AFW and COFW. The proposed method is able to produce satisfying detection results on face images with exaggerated expressions, large head poses, and partial occlusions.

cs.CV↗

arXiv2018

Joint Voxel and Coordinate Regression for Accurate 3D Facial Landmark Localization

3D face shape is more expressive and viewpoint-consistent than its 2D counterpart. However, 3D facial landmark localization in a single image is challenging due to the ambiguous nature of landmarks under 3D perspective. Existing approaches typically adopt a suboptimal two-step strategy, performing 2D landmark localization followed by depth estimation. In this paper, we propose the Joint Voxel and Coordinate Regression (JVCR) method for 3D facial landmark localization, addressing it more effectively in an end-to-end fashion. First, a compact volumetric representation is proposed to encode the per-voxel likelihood of positions being the 3D landmarks. The dimensionality of such a representation is fixed regardless of the number of target landmarks, so that the curse of dimensionality could be avoided. Then, a stacked hourglass network is adopted to estimate the volumetric representation from coarse to fine, followed by a 3D convolution network that takes the estimated volume as input and regresses 3D coordinates of the face shape. In this way, the 3D structural constraints between landmarks could be learned by the neural network in a more efficient manner. Moreover, the proposed pipeline enables end-to-end training and improves the robustness and accuracy of 3D facial landmark localization. The effectiveness of our approach is validated on the 3DFAW and AFLW2000-3D datasets. Experimental results show that the proposed method achieves state-of-the-art performance in comparison with existing methods.

cs.CV↗

arXiv2018

Global and Local Consistent Age Generative Adversarial Networks

Age progression/regression is a challenging task due to the complicated and non-linear transformation in human aging process. Many researches have shown that both global and local facial features are essential for face representation, but previous GAN based methods mainly focused on the global feature in age synthesis. To utilize both global and local facial information, we propose a Global and Local Consistent Age Generative Adversarial Network (GLCA-GAN). In our generator, a global network learns the whole facial structure and simulates the aging trend of the whole face, while three crucial facial patches are progressed or regressed by three local networks aiming at imitating subtle changes of crucial facial subregions. To preserve most of the details in age-attribute-irrelevant areas, our generator learns the residual face. Moreover, we employ an identity preserving loss to better preserve the identity information, as well as age preserving loss to enhance the accuracy of age synthesis. A pixel loss is also adopted to preserve detailed facial information of the input face. Our proposed method is evaluated on three face aging datasets, i.e., CACD dataset, Morph dataset and FG-NET dataset. Experimental results show appealing performance of the proposed method by comparing with the state-of-the-art.

cs.CV↗

arXiv2017

Learning Deep Similarity Models with Focus Ranking for Fabric Image Retrieval

Fabric image retrieval is beneficial to many applications including clothing searching, online shopping and cloth modeling. Learning pairwise image similarity is of great importance to an image retrieval task. With the resurgence of Convolutional Neural Networks (CNNs), recent works have achieved significant progresses via deep representation learning with metric embedding, which drives similar examples close to each other in a feature space, and dissimilar ones apart from each other. In this paper, we propose a novel embedding method termed focus ranking that can be easily unified into a CNN for jointly learning image representations and metrics in the context of fine-grained fabric image retrieval. Focus ranking aims to rank similar examples higher than all dissimilar ones by penalizing ranking disorders via the minimization of the overall cost attributed to similar samples being ranked below dissimilar ones. At the training stage, training samples are organized into focus ranking units for efficient optimization. We build a large-scale fabric image retrieval dataset (FIRD) with about 25,000 images of 4,300 fabrics, and test the proposed model on the FIRD dataset. Experimental results show the superiority of the proposed model over existing metric embedding models.

cs.CV↗

arXiv2017

Phase transition and intrinsic metric of the dipolar fermions in quantum Hall regime

For the fast rotating quasi-two-dimensional dipolar fermions in the quantum Hall regime, the interaction between two dipoles breaks the rotational symmetry when the dipole moment has component in the the plane via being tuned by an external field. For the anisotropic two-body interaction, we expand it in a generalized pseudopotentials (PPs). With assuming that all the dipoles are polarized in the same direction, we perform the numerical diagonalization for finite size systems on a torus. We find that the most stable fractional quantum Hall (FQH) states in the lowest Landau level (LLL) and the first Landau level (1LL) are ν = 1/3 and ν = 2 + 1/5 Laughlin state respectively in the isotropic case. While rotating the dipolar angle, these FQH states reveal a robustness and finally enter into a molecule phase in which all the particles are attracted and form a bound state. The anisotropy and the phase transition are studied by the intrinsic metric, the wave function overlap and the nematic order parameter.

cond-mat.str-el↗

arXiv2017

Deep Supervised Discrete Hashing

With the rapid growth of image and video data on the web, hashing has been extensively studied for image or video search in recent years. Benefit from recent advances in deep learning, deep hashing methods have achieved promising results for image retrieval. However, there are some limitations of previous deep hashing methods (e.g., the semantic information is not fully exploited). In this paper, we develop a deep supervised discrete hashing algorithm based on the assumption that the learned binary codes should be ideal for classification. Both the pairwise label information and the classification information are used to learn the hash codes within one stream framework. We constrain the outputs of the last layer to be binary codes directly, which is rarely investigated in deep hashing algorithm. Because of the discrete nature of hash codes, an alternating minimization method is used to optimize the objective function. Experimental results have shown that our method outperforms current state-of-the-art methods on benchmark datasets.

cs.CV↗

arXiv2017

The Interstellar Medium and Star Formation of Galactic Disks. I. ISM and GMC properties with Diffuse FUV and Cosmic Ray Backgrounds

We present a series of adaptive mesh refinement (AMR) hydrodynamic simulations of flat rotation curve galactic gas disks with a detailed treatment of the interstellar medium (ISM) physics of the atomic to molecular phase transition under the influence of diffuse FUV radiation fields and cosmic ray backgrounds. We explore the effects of different FUV intensities, including a model with a radial gradient designed to mimic the Milky Way. The effects of cosmic rays, including radial gradients in their heating and ionization rates, are also explored. The final simulations in this series achieve $4\:$pc resolution across the $\sim20\:$kpc global disk diameter, with heating and cooling followed down to temperatures of $\sim10\:$K. The disks are evolved for $300\:$Myr, which is enough time for the ISM to achieve a quasi-statistical equilibrium. In particular, the mass fraction of molecular gas stabilizes by $\sim$200 Myr. Additional global ISM properties are analysed. Giant molecular clouds (GMCs) are also identified and the statistical properties of their populations examined. GMCs are tracked as the disks evolve. GMC collisions, which may be a means of triggering star cluster formation, are counted and the rates compared with analytic models. Relatively frequent GMC collision rates are seen in these simulations and their implications for understanding GMC properties, including the driving of internal turbulence, are discussed.

astro-ph.GA↗

arXiv2017

GMC Collisions as Triggers of Star Formation. VI. Collision-Induced Turbulence

We investigate collisions between giant molecular clouds (GMCs) as potential generators of their internal turbulence. Using magnetohydrodynamic (MHD) simulations of self-gravitating, magnetized, turbulent, GMCs, we compare kinematic and dynamic properties of dense gas structures formed when such clouds collide compared to those that form in non-colliding clouds as self-gravity overwhelms decaying turbulence. We explore the nature of turbulence in these structures via distribution functions of density, velocity dispersions, virial parameters, and momentum injection. We find that the dense clumps formed from GMC collisions have higher effective Mach number, greater overall velocity dispersions, sustain near-virial equilibrium states for longer times, and are the conduit for injection of turbulent momentum into high density gas at high rates.

astro-ph.GA↗

arXiv2017

Orbital angular momentum beam instabilities in engineered nonlinear colloidal media

In this letter, we experimentally demonstrate the evolution of the optical vortex beams of different topological charges propagating in engineered nano-colloidal suspension of negative polarizability with saturable nonlinearities. Due to the high power of the incident beam, the modulation instability leads to an exponential growth of weak perturbations and thus splits the original vortex beam into a necklace beam consisting of several bright spots. The number of observed bright spots is intrinsically determined by the topological charge of the incident beam and agrees well with the predictions of our linear stability analysis and numerical simulations. Besides contributing to the fundamental science of light-matter interactions in engineered soft-matter media, this work opens new opportunities for dynamic optical manipulation and transmission of light through scattering media as well as formation of complex optical patterns and light filamentation in naturally existing colloids such as fog and clouds.

physics.optics↗

arXiv2017

Proximity-effect-induced Superconducting Gap in Topological Surface States - A Point Contact Spectroscopy Study of NbSe2/Bi2Se3 Superconductor-Topological Insulator Heterostructures

Proximity-effect-induced superconductivity was studied in epitaxial topological insulator Bi2Se3 thin films grown on superconducting NbSe2 single crystals. A point contact spectroscopy (PCS) method was used at low temperatures down to 40 mK. An induced superconducting gap in Bi2Se3 was observed in the spectra, which decreased with increasing Bi2Se3 layer thickness, consistent with the proximity effect in the bulk states of Bi2Se3 induced by NbSe2. At very low temperatures, an extra point contact feature which may correspond to a second energy gap appeared in the spectrum. For a 16 quintuple layer Bi2Se3 on NbSe2 sample, the bulk state gap value near the top surface is ~ 159 μeV, while the second gap value is ~ 120 μeV at 40 mK. The second gap value decreased with increasing Bi2Se3 layer thickness, but the ratio between the second gap and the bulk state gap remained about the same for different Bi2Se3 thicknesses. It is plausible that this is due to superconductivity in Bi2Se3 topological surface states induced through the bulk states. The two induced gaps in the PCS measurement are consistent with the three-dimensional bulk state and the two-dimensional surface state superconducting gaps observed in the angle-resolved photoemission spectroscopy (ARPES) measurement.

cond-mat.supr-con↗

arXiv2017

Preparation of CH3NH3PbI3 thin films with tens of micrometer scale at high temperature

The fabrication of high-quality organic-inorganic hybrid halide perovskite layers is the key prerequisite for the realization of high efficient photon energy harvest and electric energy conversion in their related solar cells. In this article, we report a novel fabrication technique of CH3NH3PbI3 layer based on high temperature chemical vapor reaction. CH3NH3PbI3 layers have been prepared by the reaction of PbI2 films which were deposited by pulsed laser deposition, with CH3NH3I vapor at various temperatures from 160 oC to 210 oC. X-ray diffraction patterns confirm the formation of pure phase, and photoluminescence spectra show the strong peak at around 760 nm. Scanning electron microscopy images confirm the significantly increased average grain size from nearly 1 μm at low reaction temperature of 160 oC to more than 10 μm at high reaction temperature of 200 oC. The solar cells were fabricated, and short-circuit current density of 15.75 mA/cm2, open-circuit voltage of 0.49 V and fill factor of 71.66% have been obtained.

cond-mat.mtrl-sci↗

arXiv2017

Universal properties of the FQH state from the topological entanglement entropy and disorder effects

The topological entanglement entropy (TEE) is a robust measurement of the quantum many-body state with topological order. In fractional quantum Hall (FQH) state, it has a connection to the quantum dimension of the state itself and its quasihole excitations from the conformal field theory (CFT) description. We study the entanglement entropy (EE) in the Moore-Read (MR) and Read-Rezayi (RR) FQH states. The non-Abelian quasi- hole excitation induces an extra correction of the TEE which is related to its quantum dimension. With considering the effects of the disorder, the ground state TEE is stable before the spectral gap closing and the level statistics seems to have significant change with a stronger disorder, which indicates a many-body localization (MBL) transition.

cond-mat.str-el↗

arXiv2017

Filtering Signal Processes in Molecular Multimedia Memristors

To obtain precisely controllable, robust as well as reproduceable memristor for efficient neuromorphic computing still very challenging. Molecular tailoring aims at obtaining the much more flexibly tuning plasticity has recently generated significant interest as new paradigms toward the realization of novel memristor-based synapses. Herein, inspired by the deliberate oxygen transport carried by the hemoglobin in our blood circulation, we report a novel molecular-regulated electronic/ionic semiconducting (MEIS) platform ITO/MTPP/Al2O3-x/Al with a series of metallophorphyrins (MTPPs) to delicately regulate the ionic migration for robust molecular multimedia memristors. The stable pinched hysteresis resulted from the coordination-regulated ionic migration was verified by different device structures, operation modes, as well as the characterizations of scanning transmission electron microscopy with energy-dispersive X-ray spectroscopy (STEM-EDX) and X-ray photoelectron spectroscopy (XPS). Metal coordination-dependent device parameters such as potential and depression as well as retention curves further support the correlation between the coordination and stimulating flux-dependent memristive behaviors. In the 5,10,15,20-tetraphenyl-21H,23H-porphyrin zinc(II) (ZnTPP) synapse, we implement versatile emulations, mainly including transition from short-term memory (STM) to long-term-memory (LTM), learning experience and activity-dependent learning-memory process in integrated neuromorphic configurations based on the biological Hebbian rules, and develop the fresh Spike-Amplitude-Dependent-Plasticity (SADP) with the applications of signal filtering and habituation and sensitization which are beyond the prevalent Hebbian rules.

cond-mat.mes-hall↗

arXiv2016

Learning Word Sense Embeddings from Word Sense Definitions

Word embeddings play a significant role in many modern NLP systems. Since learning one representation per word is problematic for polysemous words and homonymous words, researchers propose to use one embedding per word sense. Their approaches mainly train word sense embeddings on a corpus. In this paper, we propose to use word sense definitions to learn one embedding per word sense. Experimental results on word similarity tasks and a word sense disambiguation task show that word sense embeddings produced by our approach are of high quality.

cs.CL↗

arXiv2015

Scaling analysis of the quasiparticle tunneling in the ${\mathbb Z}_k$ parafermion states

Quasiparticle tunneling between two counter-propagating edges through point contacts could provide information on the statistics of the quasiparticles. Previous study on a disk found a scaling behavior by varying the tunneling distance. It was found that in the limit with zero tunneling distance, the Abelian quasiparticles tunneling obey the scaling analysis while the non-Abelian quasiparticles exhibit some non-trivial behaviors on the scaling exponents. Because of the limitation of disk geometry, we put the fractional quantum Hall (FQH) state on the surface of a cylinder which has a larger tunable tunneling distance than that on disk by varying the aspect ratio $γ$. We analyze the scaling behavior of the quasiholes, especially the non-Abelian quasiholes in the Read-Rezayi ${\mathbb Z}_k$ parafermion states. We aim to address the existance of the anomalous correction of the scaling parameter in the long tunneling distance.

cond-mat.str-el↗

arXiv2015

A Survey on Truth Discovery

Thanks to information explosion, data for the objects of interest can be collected from increasingly more sources. However, for the same object, there usually exist conflicts among the collected multi-source information. To tackle this challenge, truth discovery, which integrates multi-source noisy information by estimating the reliability of each source, has emerged as a hot topic. Several truth discovery methods have been proposed for various scenarios, and they have been successfully applied in diverse application domains. In this survey, we focus on providing a comprehensive overview of truth discovery methods, and summarizing them from different aspects. We also discuss some future directions of truth discovery research. We hope that this survey will promote a better understanding of the current progress on truth discovery, and offer some guidelines on how to apply these approaches in application domains.

cs.DB↗

arXiv2015

The length scale measurements of the Fractional quantum Hall state on cylinder

Once the fractional quantum Hall (FQH) state for a finite size system is put on the surface of a cylinder, the distance between the two ends with open boundary conditions can be tuned as varying the aspect ratio $γ$. It scales linearly as increasing the system size and therefore has a larger adjustable range than that on disk. The previous study of the quasi-hole tunneling amplitude on disk in Ref.~\cite{Zk2011} indicates that the tunneling amplitudes have a scaling behavior as a function of the tunneling distance and the scaling exponents are related to the scaling dimension and the charge of the transported quasiparticles. However, the scaling behaviors poorly due to the narrow range of the tunneling distance on disk. Here we systematically study the quasiparticle tunneling amplitudes of the Laughlin state in the cylinder geometry which shows a much better scaling behavior. Especially, there are some corssover behaviors at two length scales when the two open edges are close to each other. These lengths are also reflected in the bipartite entanglement and the electron Green's function as either a singularity or a crossover. These two critical length scales of the edge-edge distance, $L_x^{c_1}$ and $L_x^{c_2}$, are found to be related to the dimension reduction and back scattering point respectively.

cond-mat.str-el↗

arXiv2015

Distributed Parameter Map-Reduce

This paper describes how to convert a machine learning problem into a series of map-reduce tasks. We study logistic regression algorithm. In logistic regression algorithm, it is assumed that samples are independent and each sample is assigned a probability. Parameters are obtained by maxmizing the product of all sample probabilities. Rapid expansion of training samples brings challenges to machine learning method. Training samples are so many that they can be only stored in distributed file system and driven by map-reduce style programs. The main step of logistic regression is inference. According to map-reduce spirit, each sample makes inference through a separate map procedure. But the premise of inference is that the map procedure holds parameters for all features in the sample. In this paper, we propose Distributed Parameter Map-Reduce, in which not only samples, but also parameters are distributed in nodes of distributed filesystem. Through a series of map-reduce tasks, we assign each sample parameters for its features, make inference for the sample and update paramters of the model. The above processes are excuted looply until convergence. We test the proposed algorithm in actual hadoop production environment. Experiments show that the acceleration of the algorithm is in linear relationship with the number of cluster nodes.

cs.DC↗