arXiv subjects

Feng Huang

Publications and source records attributed to Feng Huang.

At least 37 records · Page 2Linked to original sources

arXiv2024

Single-Point Supervised High-Resolution Dynamic Network for Infrared Small Target Detection

Infrared small target detection (IRSTD) tasks are extremely challenging for two main reasons: 1) it is difficult to obtain accurate labelling information that is critical to existing methods, and 2) infrared (IR) small target information is easily lost in deep networks. To address these issues, we propose a single-point supervised high-resolution dynamic network (SSHD-Net). In contrast to existing methods, we achieve state-of-the-art (SOTA) detection performance using only single-point supervision. Specifically, we first design a high-resolution cross-feature extraction module (HCEM), that achieves bi-directional feature interaction through stepped feature cascade channels (SFCC). It balances network depth and feature resolution to maintain deep IR small-target information. Secondly, the effective integration of global and local features is achieved through the dynamic coordinate fusion module (DCFM), which enhances the anti-interference ability in complex backgrounds. In addition, we introduce the high-resolution multilevel residual module (HMRM) to enhance the semantic information extraction capability. Finally, we design the adaptive target localization detection head (ATLDH) to improve detection accuracy. Experiments on the publicly available datasets NUDT-SIRST and IRSTD-1k demonstrate the effectiveness of our method. Compared to other SOTA methods, our method can achieve better detection performance with only a single point of supervision.

cs.CV↗

arXiv2024

ZeroDDI: A Zero-Shot Drug-Drug Interaction Event Prediction Method with Semantic Enhanced Learning and Dual-Modal Uniform Alignment

Drug-drug interactions (DDIs) can result in various pharmacological changes, which can be categorized into different classes known as DDI events (DDIEs). In recent years, previously unobserved/unseen DDIEs have been emerging, posing a new classification task when unseen classes have no labelled instances in the training stage, which is formulated as a zero-shot DDIE prediction (ZS-DDIE) task. However, existing computational methods are not directly applicable to ZS-DDIE, which has two primary challenges: obtaining suitable DDIE representations and handling the class imbalance issue. To overcome these challenges, we propose a novel method named ZeroDDI for the ZS-DDIE task. Specifically, we design a biological semantic enhanced DDIE representation learning module, which emphasizes the key biological semantics and distills discriminative molecular substructure-related semantics for DDIE representation learning. Furthermore, we propose a dual-modal uniform alignment strategy to distribute drug pair representations and DDIE semantic representations uniformly in a unit sphere and align the matched ones, which can mitigate the issue of class imbalance. Extensive experiments showed that ZeroDDI surpasses the baselines and indicate that it is a promising tool for detecting unseen DDIEs. Our code has been released in https://github.com/wzy-Sarah/ZeroDDI.

cs.LG↗

arXiv2024

Heterogeneous Causal Metapath Graph Neural Network for Gene-Microbe-Disease Association Prediction

The recent focus on microbes in human medicine highlights their potential role in the genetic framework of diseases. To decode the complex interactions among genes, microbes, and diseases, computational predictions of gene-microbe-disease (GMD) associations are crucial. Existing methods primarily address gene-disease and microbe-disease associations, but the more intricate triple-wise GMD associations remain less explored. In this paper, we propose a Heterogeneous Causal Metapath Graph Neural Network (HCMGNN) to predict GMD associations. HCMGNN constructs a heterogeneous graph linking genes, microbes, and diseases through their pairwise associations, and utilizes six predefined causal metapaths to extract directed causal subgraphs, which facilitate the multi-view analysis of causal relations among three entity types. Within each subgraph, we employ a causal semantic sharing message passing network for node representation learning, coupled with an attentive fusion method to integrate these representations for predicting GMD associations. Our extensive experiments show that HCMGNN effectively predicts GMD associations and addresses association sparsity issue by enhancing the graph's semantics and structure.

cs.LG↗

arXiv2024

Efficient Meta-Learning Enabled Lightweight Multiscale Few-Shot Object Detection in Remote Sensing Images

Presently, the task of few-shot object detection (FSOD) in remote sensing images (RSIs) has become a focal point of attention. Numerous few-shot detectors, particularly those based on two-stage detectors, face challenges when dealing with the multiscale complexities inherent in RSIs. Moreover, these detectors present impractical characteristics in real-world applications, mainly due to their unwieldy model parameters when handling large amount of data. In contrast, we recognize the advantages of one-stage detectors, including high detection speed and a global receptive field. Consequently, we choose the YOLOv7 one-stage detector as a baseline and subject it to a novel meta-learning training framework. This transformation allows the detector to adeptly address FSOD tasks while capitalizing on its inherent advantage of lightweight. Additionally, we thoroughly investigate the samples generated by the meta-learning strategy and introduce a novel meta-sampling approach to retain samples produced by our designed meta-detection head. Coupled with our devised meta-cross loss, we deliberately utilize "negative samples" that are often overlooked to extract valuable knowledge from them. This approach serves to enhance detection accuracy and efficiently refine the overall meta-learning strategy. To validate the effectiveness of our proposed detector, we conducted performance comparisons with current state-of-the-art detectors using the DIOR and NWPU VHR-10.v2 datasets, yielding satisfactory results.

cs.CV↗

arXiv2024

Video Frame Interpolation for Polarization via Swin-Transformer

Video Frame Interpolation (VFI) has been extensively explored and demonstrated, yet its application to polarization remains largely unexplored. Due to the selective transmission of light by polarized filters, longer exposure times are typically required to ensure sufficient light intensity, which consequently lower the temporal sample rates. Furthermore, because polarization reflected by objects varies with shooting perspective, focusing solely on estimating pixel displacement is insufficient to accurately reconstruct the intermediate polarization. To tackle these challenges, this study proposes a multi-stage and multi-scale network called Swin-VFI based on the Swin-Transformer and introduces a tailored loss function to facilitate the network's understanding of polarization changes. To ensure the practicality of our proposed method, this study evaluates its interpolated frames in Shape from Polarization (SfP) and Human Shape Reconstruction tasks, comparing them with other state-of-the-art methods such as CAIN, FLAVR, and VFIT. Experimental results demonstrate our approach's superior reconstruction accuracy across all tasks.

cs.CV↗

arXiv2024

Multi-Level Feature Fusion Network for Lightweight Stereo Image Super-Resolution

Stereo image super-resolution utilizes the cross-view complementary information brought by the disparity effect of left and right perspective images to reconstruct higher-quality images. Cascading feature extraction modules and cross-view feature interaction modules to make use of the information from stereo images is the focus of numerous methods. However, this adds a great deal of network parameters and structural redundancy. To facilitate the application of stereo image super-resolution in downstream tasks, we propose an efficient Multi-Level Feature Fusion Network for Lightweight Stereo Image Super-Resolution (MFFSSR). Specifically, MFFSSR utilizes the Hybrid Attention Feature Extraction Block (HAFEB) to extract multi-level intra-view features. Using the channel separation strategy, HAFEB can efficiently interact with the embedded cross-view interaction module. This structural configuration can efficiently mine features inside the view while improving the efficiency of cross-view information sharing. Hence, reconstruct image details and textures more accurately. Abundant experiments demonstrate the effectiveness of MFFSSR. We achieve superior performance with fewer parameters. The source code is available at https://github.com/KarosLYX/MFFSSR.

cs.CV↗

arXiv2024

Infrared Polarization Imaging-based Non-destructive Thermography Inspection

Infrared pulse thermography non-destructive testing (NDT) method is developed based on the difference in the infrared radiation intensity emitted by defective and non-defective areas of an object. However, when the radiation intensity of the defective target is similar to that of the non-defective area of the object, the detection results are poor. To address this issue, this study investigated the polarization characteristics of the infrared radiation of different materials. Simulation results showed that the degree of infrared polarization of the object surface changed regularly with changes in thermal environment radiation. An infrared polarization imaging-based NDT method was proposed and demonstrated using specimens with four different simulated defective areas, which were designed and fabricated using four different materials. The experimental results were consistent with the simulation results, thereby proving the effectiveness of the proposed method. Compared with the infrared-radiation-intensity-based NDT method, the proposed method improved the image detail presentation and detection accuracy.

eess.IV↗

arXiv2024

Coefficient Decomposition for Spectral Graph Convolution

Spectral graph convolutional network (SGCN) is a kind of graph neural networks (GNN) based on graph signal filters, and has shown compelling expressivity for modeling graph-structured data. Most SGCNs adopt polynomial filters and learn the coefficients from the training data. Many of them focus on which polynomial basis leads to optimal expressive power and models' architecture is little discussed. In this paper, we propose a general form in terms of spectral graph convolution, where the coefficients of polynomial basis are stored in a third-order tensor. Then, we show that the convolution block in existing SGCNs can be derived by performing a certain coefficient decomposition operation on the coefficient tensor. Based on the generalized view, we develop novel spectral graph convolutions CoDeSGC-CP and -Tucker by tensor decomposition CP and Tucker on the coefficient tensor. Extensive experimental results demonstrate that the proposed convolutions achieve favorable performance improvements.

cs.LG↗

arXiv2024

X-ray polarimetric features of Gamma-ray Bursts across varied redshifts and hints for Axion-Like-Particles

Polarimetric features during the prompt phase of Gamma-ray Bursts (GRBs) have been essential for elucidating the debated emission mechanisms and gaining insight into the inner structure of GRBs. However, the potential impact of photon-Axion-Like-Particle (ALP) mixing in extragalactic magnetic fields, leading to significant modifications to the initial polarization state, has been overlooked in discussions concerning prompt phase constraints. In this work, we first examine the statistical characteristics of linear polarization degree ($Π_{L}$) in GRBs, by utilizing data from polarimetric missions focusing on sub-MeV emissions. Our analysis, conducted with a restricted sample of GRBs spanning various redshifts, reveals a diverse distribution of $Π_{L}$, which currently shows no correlation with the GRBs' spectral parameters or properties of candidate host galaxies. We then explore alternations to the initial $Π_{L}$ due to photon-ALP mixing within a domain-like structure of the intergalactic magnetic field (${\bf B}_{\rm IGM} $). With the existence of ALPs with $m_{a}$$~$$\lesssim$$~$$10^{-14}$$~$eV and $g_{aγ}~$$\simeq$$~0.5\times10^{-11}$, the mixing leads to a decrease in the polarization degree of initially fully linearly polarized photons, while it induces a certain degree of polarization to initially unpolarized photons. To ensure that the effect of mixing is small enough to be negligible, the mixing term $Δ_{aγ} \equiv 1/2\ g_{aγ} {\bf B}_{\rm IGM}$ should be less than $1.5\times 10^{-4}$ Mpc$^{-1}$. Currently, the number of GRBs with both sub-MeV polarization measurement and redshift confirmation remains very limited. Certification of redshift for GRBs with low $Π_{L}$ would further constrain the parameter space of ALPs or provide an independent means to determine the upper limit on ${\bf B}_{\rm IGM}$.

astro-ph.HE↗

arXiv2024

A Multi-Modal Contrastive Diffusion Model for Therapeutic Peptide Generation

Therapeutic peptides represent a unique class of pharmaceutical agents crucial for the treatment of human diseases. Recently, deep generative models have exhibited remarkable potential for generating therapeutic peptides, but they only utilize sequence or structure information alone, which hinders the performance in generation. In this study, we propose a Multi-Modal Contrastive Diffusion model (MMCD), fusing both sequence and structure modalities in a diffusion framework to co-generate novel peptide sequences and structures. Specifically, MMCD constructs the sequence-modal and structure-modal diffusion models, respectively, and devises a multi-modal contrastive learning strategy with intercontrastive and intra-contrastive in each diffusion timestep, aiming to capture the consistency between two modalities and boost model performance. The inter-contrastive aligns sequences and structures of peptides by maximizing the agreement of their embeddings, while the intra-contrastive differentiates therapeutic and non-therapeutic peptides by maximizing the disagreement of their sequence/structure embeddings simultaneously. The extensive experiments demonstrate that MMCD performs better than other state-of-theart deep generative methods in generating therapeutic peptides across various metrics, including antimicrobial/anticancer score, diversity, and peptide-docking.

q-bio.QM↗

arXiv2023

Extremal statistics for a one-dimensional Brownian motion with a reflective boundary

We investigate the extreme value statistics of a one-dimensional Brownian motion (with the diffusion constant $D$) during a time interval $\left[0, t \right]$ in the presence of a reflective boundary at the origin, starting from a positive position $x_0$. By deriving the survival probability of the Brownian particle without hitting an absorbing boundary at $x=M$, we obtain the distribution $P(M|x_0,t)$ of the maximum displacement $M$ and its expectation $\langle M \rangle$. In the short-time limit, i.e., $t \ll t_d$ where $t_d=x_0^2/D$ is the diffusion time from the starting position $x_0$ to the reflective boundary at the origin, the particle behaves like a free Brownian motion without any boundaries. In the long-time limit, $t \gg t_d$, $\langle M \rangle$ grows with $t$ as $\langle M \rangle \sim \sqrt{t}$, which is similar to the free Brownian motion, but the prefactor is $π/2$ times of the free Brownian motion, embodying the effect of the reflective boundary. By solving the propagator and using a path decomposition technique, we obtain the joint distribution $P(M,t_m|x_0,t)$ of $M$ and the time $t_m$ at which this maximum is achieved, from which the marginal distribution $P(t_m|x_0,t)$ is also obtained. For $t \ll t_d$, $P(t_m|x_0,t)$ looks like a U-shaped attributed to the arcsine law of free Brownian motion. For $t$ equal to or larger than order of magnitude of $t_m$, $P(t_m|x_0,t)$ deviates from the U-shaped distribution and becomes asymmetric with respect to $t/2$. Moreover, we compute the expectation $\langle t_m \rangle$ of $t_m$, and find that $\langle t_m \rangle/t$ is an increasing function of $t$. In two limiting cases, $\langle t_m \rangle/t \to 1/2$ for $t \ll t_d$ and $\langle t_m \rangle/t \to (1+2G)/4 \approx 0.708$ for $t \gg t_d$, where $G\approx0.916$ is the Catalan's constant. All the theoretical results are validated by numerical simulations.

cond-mat.stat-mech↗

arXiv2023

Testing Millisecond Pulsars as the Source of the Galactic Center Excess Gamma-Ray emission

The Galactic Center Excess (GCE) $γ$-ray emission detected with the Large Area Telescope onboard the {\it Fermi Gamma-ray Space Telescope} has been considered as a possible sign for dark matter (DM) annihilation, but other possibilities such as the millisecond pulsar (MSP) origin have also been suggested. As a spectral fitting method, constructed based on properties of $γ$-ray MSPs, has been developed, we apply this method to the study of the GCE emission for the purpose of probing the MSP origin for the GCE. A number of $\sim$1660 MSPs can provide a fit to the spectrum of the GCE emission upto $\sim$10\,GeV, but the higher energy part of the spectrum requires additional emission components. We further carry out a stacking analysis of 30--500\,GeV data for relatively nearby $γ$-ray MSPs, and the resulting flux upper limits are still lower than those of the GCE emission. We consider the single DM annihilation channel $τ^{+}τ^{-}$ or channel $b\bar{b}$,or the combination of the two for comparison, and find they generally can provide better fits than MSPs. Combination of MSPs plus a DM channel are also tested, and MSPs plus the DM channel $b\bar{b}$ can always provide better fits. Comparing this combination case to the pure DM channel $b\bar{b}$, the MSP contribution is found to be marginally needed.

astro-ph.HE↗

arXiv2023

ECMD: An Event-Centric Multisensory Driving Dataset for SLAM

Leveraging multiple sensors enhances complex environmental perception and increases resilience to varying luminance conditions and high-speed motion patterns, achieving precise localization and mapping. This paper proposes, ECMD, an event-centric multisensory dataset containing 81 sequences and covering over 200 km of various challenging driving scenarios including high-speed motion, repetitive scenarios, dynamic objects, etc. ECMD provides data from two sets of stereo event cameras with different resolutions (640*480, 346*260), stereo industrial cameras, an infrared camera, a top-installed mechanical LiDAR with two slanted LiDARs, two consumer-level GNSS receivers, and an onboard IMU. Meanwhile, the ground-truth of the vehicle was obtained using a centimeter-level high-accuracy GNSS-RTK/INS navigation system. All sensors are well-calibrated and temporally synchronized at the hardware level, with recording data simultaneously. We additionally evaluate several state-of-the-art SLAM algorithms for benchmarking visual and LiDAR SLAM and identifying their limitations. The dataset is available at https://arclab-hku.github.io/ecmd/.

cs.RO↗

arXiv2023

Comparative study of flow fluctuations in ruptured and unruptured intracranial aneurysms: A lattice Boltzmann study

Flow fluctuations have recently emerged as a promising hemodynamic metric for understanding the rupture risk of intracranial aneurysms. Several investigations have reported in the literature corresponding flow instabilities using established computational fluid dynamics tools. In this study, the occurrence of flow fluctuations is investigated using either Newtonian or non-Newtonian fluid models in patient-specific intracranial aneurysms using high-resolution lattice Boltzmann method simulations. Flow instabilities are quantified by computing power spectral density, proper orthogonal decomposition and spectral entropy, and fluctuating kinetic energy of velocity fluctuations. Furthermore, these hemodynamic parameters are compared between the ruptured and unruptured aneurysms. Our simulations reveal that the pulsatile inflow through the neck in a ruptured aneurysm is subject to a hydrodynamic instability leading to high-frequency fluctuations around the rupture position throughout the entire cardiac cycle. At other locations, the flow instability is only observed during the deceleration phase; typically, the fluctuations begin there just after peak systole, gradually decay, and the flow returns to its original, laminar pulsatile state during diastole. In the unruptured aneurysm, there is only minimal difference between Newtonian and non-Newtonian results. In the ruptured case, using the non-Newtonian model leads to a considerable increase of the fluctuations within the aneurysm sac.

physics.flu-dyn↗

arXiv2023

Constraints on dark matter annihilation from the FAST observation of the Coma Berenices dwarf galaxy

The Galactic center $γ$-ray excess, detected by the Fermi-LAT, is a very attractive tentative signal from dark matter annihilation. Searching for associated synchrotron emissions can test the dark matter interpretation for this excess. We geared the Five-hundred-meter Aperture Spherical radio Telescope (FAST) towards Coma Berenices (a dwarf Spheroidal galaxy) for 2-hours of observation, and found no significant continuum radio emission, which could { put constraints on the dark matter annihilation cross section}, from our target. We set very stringent annihilation cross-section constraints, with roughly an order of magnitude improvement over a large range of masses compared with previous radio searches. The dark matter scenario for the Galactic center $γ$-ray excess is in tension with the FAST observation for reasonable choices of astrophysical factors. But considering the large uncertainty on astrophysical parameters, such as the magnetic field, the diffusion coefficient, and the diffusion radius, and on DM halo parameters, the dark matter interpretation for the excess could still survive. Further radio observations by the FAST and other radio telescopes may reveal more about the dark matter properties.

astro-ph.HE↗

arXiv2022

3D LiDAR Aided GNSS NLOS Mitigation for Reliable GNSS-RTK Positioning in Urban Canyons

GNSS and LiDAR odometry are complementary as they provide absolute and relative positioning, respectively. Their integration in a loosely-coupled manner is straightforward but is challenged in urban canyons due to the GNSS signal reflections. Recent proposed 3D LiDAR-aided (3DLA) GNSS methods employ the point cloud map to identify the non-line-of-sight (NLOS) reception of GNSS signals. This facilitates the GNSS receiver to obtain improved urban positioning but not achieve a sub-meter level. GNSS real-time kinematics (RTK) uses carrier phase measurements to obtain decimeter-level positioning. In urban areas, the GNSS RTK is not only challenged by multipath and NLOS-affected measurement but also suffers from signal blockage by the building. The latter will impose a challenge in solving the ambiguity within the carrier phase measurements. In the other words, the model observability of the ambiguity resolution (AR) is greatly decreased. This paper proposes to generate virtual satellite (VS) measurements using the selected LiDAR landmarks from the accumulated 3D point cloud maps (PCM). These LiDAR-PCM-made VS measurements are tightly-coupled with GNSS pseudorange and carrier phase measurements. Thus, the VS measurements can provide complementary constraints, meaning providing low-elevation-angle measurements in the across-street directions. The implementation is done using factor graph optimization to solve an accurate float solution of the ambiguity before it is fed into LAMBDA. The effectiveness of the proposed method has been validated by the evaluation conducted on our recently open-sourced challenging dataset, UrbanNav. The result shows the fix rate of the proposed 3DLA GNSS RTK is about 30% while the conventional GNSS-RTK only achieves about 14%. In addition, the proposed method achieves sub-meter positioning accuracy in most of the data collected in challenging urban areas.

cs.RO↗

arXiv2022

Non-equilibrium random walks on multiplex networks

We introduce a non-equilibrium discrete-time random walk model on multiplex networks, in which at each time step the walker first undergoes a random jump between neighboring nodes in the same layer, and then tries to hop from one node to one of its replicas in another layer. We derive the so-called supra-Markov matrix that governs the evolution of the occupation probability of the walker. The occupation probability at stationarity is different from the weighted average over the counterparts on each layer, unless the transition probabilities between layers vanish. However, they are approximately equal when the transition probabilities between layers are very small, which is given by the first-order degenerate perturbation theory. Moreover, we compute the mean first passage time (MFPT) and the graph MFPT (GrMFPT) that is the average of the MFPT over all pairs of distinct nodes. Interestingly, we find that the GrMFPT can be smaller than that of any layer taken in isolation. The result embodies the advantage of global search on multiplex networks.

cond-mat.stat-mech↗

arXiv2022

Dark matter admixed neutron star properties in the light of X-ray pulse profile observations

The distribution of the dark matter (DM) in DM-admixed-neutron stars (DANSs) is supposed to be either a dense dark core or an extended dark halo, which is subject to the DM fraction of DANS ($f_χ$) and the DM properties, such as the mass ($m_χ$) and the strength of the self-interaction ($y$). In this paper, we perform an in-depth analysis of the formation criterion for dark core/dark halo and point out that the relative distribution of these two components is essentially determined by the ratio of the central enthalpy of the DM component to that of the baryonic matter component inside DANSs. For the critical case where the radii of DM and baryonic matter are the same, we further derive an analytical formula to describe the dependence of $f^{\rm crit}_χ$ on $m_χ$ and $y$ for given DANS mass. The relative distribution of the two components in DANSs can lead to different observational effects. We here focus on the modification of the pulsar pulse profile due to the extra light-bending effect in the case of a dark-halo existence and conduct the first investigation of the dark-halo effects on the pulse profile. We find that the peak flux deviation is strongly dependent on the ratio of the halo mass to the radius of the DM component. Lastly, we perform Bayesian parameter estimation on the DM particle properties based on the recent X-ray observations of PSR J0030+0451 and PSR J0740+6620 by the Neutron Star Interior Composition Explorer.

astro-ph.HE↗