arXiv Science⌕ Search

arXiv subjects

Yuhan Wang

Publications and source records attributed to Yuhan Wang.

176 records · Page 10Linked to original sources
arXiv2021

HifiFace: 3D Shape and Semantic Prior Guided High Fidelity Face Swapping

In this work, we propose a high fidelity face swapping method, called HifiFace, which can well preserve the face shape of the source face and generate photo-realistic results. Unlike other existing face swapping works that only use face recognition model to keep the identity similarity, we propose 3D shape-aware identity to control the face shape with the geometric supervision from 3DMM and 3D face reconstruction method. Meanwhile, we introduce the Semantic Facial Fusion module to optimize the combination of encoder and decoder features and make adaptive blending, which makes the results more photo-realistic. Extensive experiments on faces in the wild demonstrate that our method can preserve better identity, especially on the face shape, and can generate more photo-realistic results than previous state-of-the-art methods.

cs.CV↗
arXiv2021

Controlling relaxation dynamics of excitonic states in monolayer transition metal dichalcogenides WS2 through interface engineering

Transition metal dichalcogenides (TMDs) are known to support complex excitonic states. Revealing the differences in relaxation dynamics among different excitonic species and elucidating the transition dynamics between them may provide important guidelines for designing TMD-based excitonic devices. Combining photoluminescence (PL) and reflectance contrast measurements with ultrafast pump-probe spectroscopy under cryogenic temperatures, we herein study the relaxation dynamics of neutral and charged excitons in a back-gate-controlled monolayer device. Pump-probe results reveal quite different relaxation dynamics of excitonic states under different interfacial conditions: while neutral excitons experience much longer lifetime than trions in monolayer WS2, the opposite is true in the WS2/h-BN heterostructure. It is found that the insertion of h-BN layer between the TMD monolayer and the substrate has a great influence on the lifetimes of different excitonic states. The h-BN flakes can not only screen the effects of impurities and defects at the interface, but also help establish a non-radiative transition from neutral excitons to trions to be the dominant relaxation pathway, under cryogenic temperature. Our findings highlight the important role interface may play in governing the transient properties of carriers in 2D semiconductors, and may also have implications for designing light-emitting and photo-detecting devices based on TMDs.

cond-mat.mes-hall↗
arXiv2021

In situ Performance of the Low Frequency Arrayfor Advanced ACTPol

The Advanced Atacama Cosmology Telescope Polarimeter (AdvACT) \cite{thornton} is an upgrade for the Atacama Cosmology Telescope using Transition Edge Sensor (TES) detector arrays to measure cosmic microwave background (CMB) temperature and polarization anisotropies in multiple frequencies. The low frequency (LF) array was deployed early 2020. It consists of 292 TES bolometers observing in two bands centered at 27 GHz and 39 GHz. At these frequencies, it is sensitive to synchrotron radiation from our galaxy as well as to the CMB, and complements the AdvACT arrays operating at 90, 150 and 230 GHz. We present the initial LF array on-site characterization, including the time constant, optical efficiency and array sensitivity.

astro-ph.IM↗
arXiv2020

The Simons Observatory: the Large Aperture Telescope Receiver (LATR) Integration and Validation Results

The Simons Observatory (SO) will observe the cosmic microwave background (CMB) from Cerro Toco in the Atacama Desert of Chile. The observatory consists of three 0.5 m Small Aperture Telescopes (SATs) and one 6 m Large Aperture Telescope (LAT), covering six frequency bands centering around 30, 40, 90, 150, 230, and 280 GHz. The SO observations will transform the understanding of our universe by characterizing the properties of the early universe, measuring the number of relativistic species and the mass of neutrinos, improving our understanding of galaxy evolution, and constraining the properties of cosmic reionization. As a critical instrument, the Large Aperture Telescope Receiver (LATR) is designed to cool $\sim$ 60,000 transition-edge sensors (TES) to $<$ 100 mK on a 1.7 m diameter focal plane. The unprecedented scale of the LATR drives a complex design. In this paper, we will first provide an overview of the LATR design. Integration and validation of the LATR design are discussed in detail, including mechanical strength, optical alignment, and cryogenic performance of the five cryogenic stages (80 K, 40 K, 4 K, 1 K, and 100 mK). We will also discuss the microwave-multiplexing ($μ$Mux) readout system implemented in the LATR and demonstrate the operation of dark prototype TES bolometers. The $μ$Mux readout technology enables one coaxial loop to read out $\mathcal{O}(10^3)$ TES detectors. Its implementation within the LATR serves as a critical validation for the complex RF chain design. The successful validation of the LATR performance is not only a critical milestone within the Simons Observatory, it also provides a valuable reference for other experiments, e.g. CCAT-prime and CMB-S4.

astro-ph.IM↗
arXiv2020

Effective and Sparse Count-Sketch via k-means clustering

Count-sketch is a popular matrix sketching algorithm that can produce a sketch of an input data matrix X in O(nnz(X))time where nnz(X) denotes the number of non-zero entries in X. The sketched matrix will be much smaller than X while preserving most of its properties. Therefore, count-sketch is widely used for addressing high-dimensionality challenge in machine learning. However, there are two main limitations of count-sketch: (1) The sketching matrix used count-sketch is generated randomly which does not consider any intrinsic data properties of X. This data-oblivious matrix sketching method could produce a bad sketched matrix which will result in low accuracy for subsequent machine learning tasks (e.g.classification); (2) For highly sparse input data, count-sketch could produce a dense sketched data matrix. This dense sketch matrix could make the subsequent machine learning tasks more computationally expensive than on the original sparse data X. To address these two limitations, we first show an interesting connection between count-sketch and k-means clustering by analyzing the reconstruction error of the count-sketch method. Based on our analysis, we propose to reduce the reconstruction error of count-sketch by using k-means clustering algorithm to obtain the low-dimensional sketched matrix. In addition, we propose to solve k-mean clustering using gradient descent with -L1 ball projection to produce a sparse sketched matrix. Our experimental results based on six real-life classification datasets have demonstrated that our proposed method achieves higher accuracy than the original count-sketch and other popular matrix sketching algorithms. Our results also demonstrate that our method produces a sparser sketched data matrix than other methods and therefore the prediction cost of our method will be smaller than other matrix sketching methods.

cs.LG↗
arXiv2020

CurricularFace: Adaptive Curriculum Learning Loss for Deep Face Recognition

As an emerging topic in face recognition, designing margin-based loss functions can increase the feature margin between different classes for enhanced discriminability. More recently, the idea of mining-based strategies is adopted to emphasize the misclassified samples, achieving promising results. However, during the entire training process, the prior methods either do not explicitly emphasize the sample based on its importance that renders the hard samples not fully exploited; or explicitly emphasize the effects of semi-hard/hard samples even at the early training stage that may lead to convergence issue. In this work, we propose a novel Adaptive Curriculum Learning loss (CurricularFace) that embeds the idea of curriculum learning into the loss function to achieve a novel training strategy for deep face recognition, which mainly addresses easy samples in the early training stage and hard ones in the later stage. Specifically, our CurricularFace adaptively adjusts the relative importance of easy and hard samples during different training stages. In each stage, different samples are assigned with different importance according to their corresponding difficultness. Extensive experimental results on popular benchmarks demonstrate the superiority of our CurricularFace over the state-of-the-art competitors.

cs.CV↗
arXiv2020

Intelligent Home 3D: Automatic 3D-House Design from Linguistic Descriptions Only

Home design is a complex task that normally requires architects to finish with their professional skills and tools. It will be fascinating that if one can produce a house plan intuitively without knowing much knowledge about home design and experience of using complex designing tools, for example, via natural language. In this paper, we formulate it as a language conditioned visual content generation problem that is further divided into a floor plan generation and an interior texture (such as floor and wall) synthesis task. The only control signal of the generation process is the linguistic expression given by users that describe the house details. To this end, we propose a House Plan Generative Model (HPGM) that first translates the language input to a structural graph representation and then predicts the layout of rooms with a Graph Conditioned Layout Prediction Network (GC LPN) and generates the interior texture with a Language Conditioned Texture GAN (LCT-GAN). With some post-processing, the final product of this task is a 3D house model. To train and evaluate our model, we build the first Text-to-3D House Model dataset.

cs.CV↗
arXiv2019

Enhancing the Three-Photon Luminesce via the Resonance in Lead Halide Perovskite Metasurfaces

Lead halide perovskites (MAPbX3) have emerged as promising materials for photovoltaic and optoelectronic devices. However, their exceptional nonlinear properties have not been fully exploited in nanophotonics yet. Herein we fabricate MAPbX3 perovskite metasurfaces and explore their internal nonlinear processes. While both of third-order harmonic generation (THG) and three-photon luminescence are generated, the latter one is less affected by the material loss and has been significantly enhanced by a factor of 60. The corresponding simulation reveals that the improvement is caused by the resonant enhancement of incident laser in perovskite metasurface. Interestingly, such kind of resonance-enhanced three-photon luminescence holds true for metasurfaces with a small period number of 4, enabling new applications of perovskite metasurface in high-resolution nonlinear color nanoprinting and optical encoding. The encoded information "NANO" is visible only when the incident laser is on-resonance. The off-resonance pumping and the single-photon excitation just produce a uniform dark or photoluminescence background.

physics.app-ph↗
arXiv2019

Lead Halide Perovskite based Dynamic Metasurfaces

Lead halide perovskites (MAPbX3) are known to have high refractive index and controllable bandgap, making them attractive for all-dielectric and tunable metasurfaces. Till now, perovskite metasurfaces have only been used in structural colors. More interesting meta-devices with 2π phase control are still absent. Here we experimentally demonstrate the MAPbX3 perovskite based metasurfaces with a complete control of phase shift in a reflection mode. By utilizing MAPbBr3 cut-wires as meta-atoms on a ground metal film, we find that the MAPbBr3 perovskite metasurface can produce full phase control from 0 to 2π and high reflection efficiency simultaneously. Consequently, high-efficiency polarization conversion, anomalous reflection and meta-hologram have been successfully produced. Most interestingly, the bandgap of MAPbX3 perovskite can be post-synthetically and reversibly tuned via anion exchange, providing a new approach to dynamically control of the all-dielectric meta-devices with novel function such as anomalous reflection and hologram et al.

physics.app-ph↗
arXiv2018

Ultrafast photocarrier recombination dynamics in black phosphorus-molybdenum disulfide (BP/MoS2) heterostructure

Van der Waals (vdW) heterostructures constructed with two-dimensional (2D) materials have attracted great interests, due to their fascinating properties and potential for novel applications. While earlier efforts have advanced the understanding of the ultrafast cross-layer charge transfer process in 2D heterostructures, mechanisms for the interfacial photocarrier recombination remain, to a large extent, unclear. Here, we investigate a heterostructure comprised of black phosphorus (BP) and molybdenum disulfide (MoS2), with a type-II band alignment. Interestingly, it is found that the photo-generated electrons in MoS2 (transferred from BP) exhibit an ultrafast lifetime of about 5 ps, significantly shorter than those of the constituent materials. By corroborating with the relaxation of photo-excited holes in BP, it is revealed that the ultrafast time constant is as a result of efficient Langevin recombination, where the high hole mobility of BP facilitates a large recombination coefficient (approximately 2x10^-10 m^2/s). In addition, broadband transient absorption spectroscopy confirms that the hot electrons transferred to MoS2 distribute over a broad energy range following an ultrafast thermalization. The rate of the interlayer Langevin recombination is found to exhibit no energy level dependence. Our findings provide essential insights into the fundamental photo-physics in type-II 2D heterostructures, and also provide useful guidelines for customizing photocarrier lifetimes of BP for high-speed photo-sensitive devices.

cond-mat.mes-hall↗
arXiv2018

Orientational migration of halide ion in lead halide perovskites

With the rapid evolution of lead halide perovskite solar cells, there are considerable evidences that the mobile ions strongly impact their properties. However, except the activation energy, the other properties of ionic migration are mostly explored with first principle calculations. Here we investigate the migration of halide ion (vacancy) in lead bromide perovskite single crystalline microplates. In contrast to conventional studies with external bias voltages, we introduce built-in potentials into the microplates via low-energy electron beams and study the corresponding ionic migrations. By aligning the potentials along particular crystal planes, the reduction of Pb2+ to Pb0 has been observed at the un-irradiated center of each cell, demonstrating that the migration of halide ion has preferred directions for the first time. Our observations are consistent with the recent theoretical models and shall open up a new route to improve the performances of perovskite solar cells.

cond-mat.mtrl-sci↗
arXiv2017

Automatic Structural Scene Digitalization

In this paper, we present an automatic system for the analysis and labeling of structural scenes, floor plan drawings in Computer-aided Design (CAD) format. The proposed system applies a fusion strategy to detect and recognize various components of CAD floor plans, such as walls, doors, windows and other ambiguous assets. Technically, a general rule-based filter parsing method is fist adopted to extract effective information from the original floor plan. Then, an image-processing based recovery method is employed to correct information extracted in the first step. Our proposed method is fully automatic and real-time. Such analysis system provides high accuracy and is also evaluated on a public website that, on average, archives more than ten thousands effective uses per day and reaches a relatively high satisfaction rate.

cs.GR↗
arXiv2017

Reflectance dependence of polytetrafluoroethylene on thickness for xenon scintillation light

Many rare event searches including dark matter direct detection and neutrinoless double beta decay experiments take advantage of the high VUV reflective surfaces made from polytetrafluoroethylene (PTFE) reflector materials to achieve high light collection efficiency in their detectors. As the detectors have grown in size over the past decade, there has also been an increased need for ever thinner detector walls without significant loss in reflectance to reduce dead volumes around active noble liquids, outgassing, and potential backgrounds. We report on the experimental results to measure the dependence of the reflectance on thickness of two PTFE samples at wavelengths near 178 nm. No change in reflectance was observed as the thickness of a cylindrically shaped PTFE vessel immersed in liquid xenon was varied between 1 mm and 9.5 mm.

physics.ins-det↗