Searcharxiv⌕ Search

arXiv subjects

Jun Li

Publications and source records attributed to Jun Li.

At least 271 records · Page 15Linked to original sources
arXiv2025

Variable Radiance Field for Real-World Category-Specific Reconstruction from Single Image

Reconstructing category-specific objects using Neural Radiance Field (NeRF) from a single image is a promising yet challenging task. Existing approaches predominantly rely on projection-based feature retrieval to associate 3D points in the radiance field with local image features from the reference image. However, this process is computationally expensive, dependent on known camera intrinsics, and susceptible to occlusions. To address these limitations, we propose Variable Radiance Field (VRF), a novel framework capable of efficiently reconstructing category-specific objects without requiring known camera intrinsics and demonstrating robustness against occlusions. First, we replace the local feature retrieval with global latent representations, generated through a single feed-forward pass, which improves efficiency and eliminates reliance on camera intrinsics. Second, to tackle coordinate inconsistencies inherent in real-world dataset, we define a canonical space by introducing a learnable, category-specific shape template and explicitly aligning each training object to this template using a learnable 3D transformation. This approach also reduces the complexity of geometry prediction to modeling deformations from the template to individual instances. Finally, we employ a hyper-network-based method for efficient NeRF creation and enhance the reconstruction performance through a contrastive learning-based pretraining strategy. Evaluations on the CO3D dataset demonstrate that VRF achieves state-of-the-art performance in both reconstruction quality and computational efficiency.

cs.CV↗
arXiv2025

Reading Your Heart: Learning ECG Words and Sentences via Pre-training ECG Language Model

Electrocardiogram (ECG) is essential for the clinical diagnosis of arrhythmias and other heart diseases, but deep learning methods based on ECG often face limitations due to the need for high-quality annotations. Although previous ECG self-supervised learning (eSSL) methods have made significant progress in representation learning from unannotated ECG data, they typically treat ECG signals as ordinary time-series data, segmenting the signals using fixed-size and fixed-step time windows, which often ignore the form and rhythm characteristics and latent semantic relationships in ECG signals. In this work, we introduce a novel perspective on ECG signals, treating heartbeats as words and rhythms as sentences. Based on this perspective, we first designed the QRS-Tokenizer, which generates semantically meaningful ECG sentences from the raw ECG signals. Building on these, we then propose HeartLang, a novel self-supervised learning framework for ECG language processing, learning general representations at form and rhythm levels. Additionally, we construct the largest heartbeat-based ECG vocabulary to date, which will further advance the development of ECG language processing. We evaluated HeartLang across six public ECG datasets, where it demonstrated robust competitiveness against other eSSL methods. Our data and code are publicly available at https://github.com/PKUDigitalHealth/HeartLang.

cs.LG↗
arXiv2025

Dynamic Appearance Particle Neural Radiance Field

Neural Radiance Fields (NeRFs) have shown great potential in modeling 3D scenes. Dynamic NeRFs extend this model by capturing time-varying elements, typically using deformation fields. The existing dynamic NeRFs employ a similar Eulerian representation for both light radiance and deformation fields. This leads to a close coupling of appearance and motion and lacks a physical interpretation. In this work, we propose Dynamic Appearance Particle Neural Radiance Field (DAP-NeRF), which introduces particle-based representation to model the motions of visual elements in a dynamic 3D scene. DAP-NeRF consists of the superposition of a static field and a dynamic field. The dynamic field is quantized as a collection of appearance particles, which carries the visual information of a small dynamic element in the scene and is equipped with a motion model. All components, including the static field, the visual features and the motion models of particles, are learned from monocular videos without any prior geometric knowledge of the scene. We develop an efficient computational framework for the particle-based model. We also construct a new dataset to evaluate motion modeling. Experimental results show that DAP-NeRF is an effective technique to capture not only the appearance but also the physically meaningful motions in a 3D dynamic scene. Code is available at: https://github.com/Cenbylin/DAP-NeRF.

cs.CV↗
arXiv2025

Normal Transformer: Extracting Surface Geometry from LiDAR Points Enhanced by Visual Semantics

High-quality surface normal can help improve geometry estimation in problems faced by autonomous vehicles, such as collision avoidance and occlusion inference. While a considerable volume of literature focuses on densely scanned indoor scenarios, normal estimation during autonomous driving remains an intricate problem due to the sparse, non-uniform, and noisy nature of real-world LiDAR scans. In this paper, we introduce a multi-modal technique that leverages 3D point clouds and 2D colour images obtained from LiDAR and camera sensors for surface normal estimation. We present the Hybrid Geometric Transformer (HGT), a novel transformer-based neural network architecture that proficiently fuses visual semantic and 3D geometric information. Furthermore, we developed an effective learning strategy for the multi-modal data. Experimental results demonstrate the superior effectiveness of our information fusion approach compared to existing methods. It has also been verified that the proposed model can learn from a simulated 3D environment that mimics a traffic scene. The learned geometric knowledge is transferable and can be applied to real-world 3D scenes in the KITTI dataset. Further tasks built upon the estimated normal vectors in the KITTI dataset show that the proposed estimator has an advantage over existing methods.

cs.CV↗
arXiv2025

Learning Inverse Laplacian Pyramid for Progressive Depth Completion

Depth completion endeavors to reconstruct a dense depth map from sparse depth measurements, leveraging the information provided by a corresponding color image. Existing approaches mostly hinge on single-scale propagation strategies that iteratively ameliorate initial coarse depth estimates through pixel-level message passing. Despite their commendable outcomes, these techniques are frequently hampered by computational inefficiencies and a limited grasp of scene context. To circumvent these challenges, we introduce LP-Net, an innovative framework that implements a multi-scale, progressive prediction paradigm based on Laplacian Pyramid decomposition. Diverging from propagation-based approaches, LP-Net initiates with a rudimentary, low-resolution depth prediction to encapsulate the global scene context, subsequently refining this through successive upsampling and the reinstatement of high-frequency details at incremental scales. We have developed two novel modules to bolster this strategy: 1) the Multi-path Feature Pyramid module, which segregates feature maps into discrete pathways, employing multi-scale transformations to amalgamate comprehensive spatial information, and 2) the Selective Depth Filtering module, which dynamically learns to apply both smoothness and sharpness filters to judiciously mitigate noise while accentuating intricate details. By integrating these advancements, LP-Net not only secures state-of-the-art (SOTA) performance across both outdoor and indoor benchmarks such as KITTI, NYUv2, and TOFDC, but also demonstrates superior computational efficiency. At the time of submission, LP-Net ranks 1st among all peer-reviewed methods on the official KITTI leaderboard.

cs.CV↗
arXiv2025

A one-dimensional mixing model for the impact of ablative Rayleigh-Taylor instability on compression dynamics

A one-dimensional mixing model, incorporating the effects of laser ablation and initial perturbations, is developed to study the influence of ablative Rayleigh-Taylor instability on compression dynamics. The length of the mixing region is determined with the buoyancy-drag model[arXiv:2411.12392v2 (2024)]. The mixing effect on laser ablation is mainly described with an additional heat source which depends on turbulent kinetic energy and initial perturbation level through a free multiplier. The model is integrated into a one-dimensional radiation hydrodynamics code and validated against two-dimensional planar simulations. The further application of our model to spherical implosion simulations reveals that the model can give reasonable predictions of implosion degradation due to mixing, such as lowered shell compression, reduced stagnation pressure, and decreased areal density, etc. It is found that the time interval between the convergence of the main shock and stagnation may offer an estimate of mixing level in single-shot experiments.

physics.flu-dyn↗
arXiv2025

Harmonious Group Choreography with Trajectory-Controllable Diffusion

Creating group choreography from music is crucial in cultural entertainment and virtual reality, with a focus on generating harmonious movements. Despite growing interest, recent approaches often struggle with two major challenges: multi-dancer collisions and single-dancer foot sliding. To address these challenges, we propose a Trajectory-Controllable Diffusion (TCDiff) framework, which leverages non-overlapping trajectories to ensure coherent and aesthetically pleasing dance movements. To mitigate collisions, we introduce a Dance-Trajectory Navigator that generates collision-free trajectories for multiple dancers, utilizing a distance-consistency loss to maintain optimal spacing. Furthermore, to reduce foot sliding, we present a footwork adaptor that adjusts trajectory displacement between frames, supported by a relative forward-kinematic loss to further reinforce the correlation between movements and trajectories. Experiments demonstrate our method's superiority.

cs.CV↗
arXiv2025

Decision Transformers for RIS-Assisted Systems with Diffusion Model-Based Channel Acquisition

Reconfigurable intelligent surfaces (RISs) have been recognized as a revolutionary technology for future wireless networks. However, RIS-assisted communications have to continuously tune phase-shifts relying on accurate channel state information (CSI) that is generally difficult to obtain due to the large number of RIS channels. The joint design of CSI acquisition and subsection RIS phase-shifts remains a significant challenge in dynamic environments. In this paper, we propose a diffusion-enhanced decision Transformer (DEDT) framework consisting of a diffusion model (DM) designed for efficient CSI acquisition and a decision Transformer (DT) utilized for phase-shift optimizations. Specifically, we first propose a novel DM mechanism, i.e., conditional imputation based on denoising diffusion probabilistic model, for rapidly acquiring real-time full CSI by exploiting the spatial correlations inherent in wireless channels. Then, we optimize beamforming schemes based on the DT architecture, which pre-trains on historical environments to establish a robust policy model. Next, we incorporate a fine-tuning mechanism to ensure rapid beamforming adaptation to new environments, eliminating the retraining process that is imperative in conventional reinforcement learning (RL) methods. Simulation results demonstrate that DEDT can enhance efficiency and adaptability of RIS-aided communications with fluctuating channel conditions compared to state-of-the-art RL methods.

eess.SP↗
arXiv2025

Experimental Realization of Criticality-Enhanced Global Quantum Sensing via Non-Equilibrium Dynamics

Quantum critical systems offer promising advancements in quantum sensing and metrology, yet face limitations like critical slowing down and a restricted criticality-enhanced region. Here, we introduce a critical sensing scheme that mitigate critical slowing down by leveraging the non-equilibrium dynamics of a perturbed Ising spin model, coupled with an adaptive strategy to enlarge its sensing interval. We validate the proposed scheme on a superconducting quantum processor and demonstrate that our scheme achieves a Heisenberg scaling with respect to the encoding duration. Additionally, the adaptive strategy tunes the model to operate near its critical point with limited prior information about the parameter, enabling what is known as global sensing. Our work showcases the metrological applications empowered by non-equilibrium critical dynamics and hence opens up a pathway for devising critical quantum sensors.

quant-ph↗
arXiv2025

SFADNet: Spatio-temporal Fused Graph based on Attention Decoupling Network for Traffic Prediction

In recent years, traffic flow prediction has played a crucial role in the management of intelligent transportation systems. However, traditional prediction methods are often limited by static spatial modeling, making it difficult to accurately capture the dynamic and complex relationships between time and space, thereby affecting prediction accuracy. This paper proposes an innovative traffic flow prediction network, SFADNet, which categorizes traffic flow into multiple traffic patterns based on temporal and spatial feature matrices. For each pattern, we construct an independent adaptive spatio-temporal fusion graph based on a cross-attention mechanism, employing residual graph convolution modules and time series modules to better capture dynamic spatio-temporal relationships under different fine-grained traffic patterns. Extensive experimental results demonstrate that SFADNet outperforms current state-of-the-art baselines across four large-scale datasets.

cs.LG↗
arXiv2025

Contribution of Unresolved Sources to Diffuse Gamma-Ray Emission from the Galactic Plane

The diffuse gamma-ray emission from the Milky Way serves as a crucial probe for understanding the propagation and interactions of cosmic rays within our galaxy. The Galactic diffuse gamma-ray emission between 10 TeV and 1 PeV has been recently measured by the square kilometer array (KM2A) of the Large High Altitude Air Shower Observatory (LHAASO). The flux is higher than predicted for cosmic rays interacting with the interstellar medium. In this work, we utilize a non-parametric method to derive the source count distribution using the published first LHAASO source catalog. Based on this distribution, we calculate the contribution of unresolved sources to the diffuse emission measured by KM2A. When comparing our results to the measured diffuse gamma-ray emission, we demonstrate that for the outer Galactic region, the contributions from unresolved sources and those predicted by models are roughly consistent with experimental observations within the uncertainty. However, for the inner Galactic region, additional components are required to account for the observed data.

astro-ph.HE↗
arXiv2025

A buoyancy-drag model with a time-varying drag coefficient for evaluating bubble front penetration depth

To evaluate and control bubble front penetration depth ${{h}_{B}}$ induced by ablative Rayleigh-Taylor instability (ARTI) from a weakly nonlinear phase to a self-similar phase, we first propose an improved buoyancy-drag (BD) model with a time-varying drag coefficient. The coefficient incorporates the influence of multiple physical mechanisms, including non-steady ablation, preheating, and other mechanisms during this phase. The model is validated through simulations under various conditions, demonstrating improved accuracy compared to the classical BD model and the self-similar growth. Furthermore, the model suggests controlling ${{h}_{B}}$ by suppressing the "most dangerous mode", which is influenced by initial perturbations and ablative acceleration history, thus offering novel insights for target manufacturing and pulse optimization near the ignition threshold.

physics.flu-dyn↗
arXiv2024

Throughput Maximization for UAV-enabled Integrated Periodic Sensing and Communication

Unmanned aerial vehicle (UAV) is expected to revolutionize the existing integrated sensing and communication (ISAC) system and promise a more flexible joint design. Nevertheless, the existing works on ISAC mainly focus on exploring the performance of both functionalities simultaneously during the entire considered period, which may ignore the practical asymmetric sensing and communication requirements. In particular, always forcing sensing along with communication may make it is harder to balance between these two functionalities due to shared spectrum resources and limited transmit power. To address this issue, we propose a new integrated periodic sensing and communication mechanism for the UAV-enabled ISAC system to provide a more flexible trade-off between two integrated functionalities. Specifically, the system achievable rate is maximized via jointly optimizing UAV trajectory, user association, target sensing selection, and transmit beamforming, while meeting the sensing frequency and beam pattern gain requirement for the given targets. Despite that this problem is highly non-convex and involves closely coupled integer variables, we derive the closed-form optimal beamforming vector to dramatically reduce the complexity of beamforming design, and present a tight lower bound of the achievable rate to facilitate UAV trajectory design. Based on the above results, we propose a penalty-based algorithm to efficiently solve the considered problem. The optimal achievable rate and the optimal UAV location are analyzed under a special case of infinity number of antennas. Furthermore, we prove the structural symmetry between the optimal solutions in different ISAC frames without location constraints and propose an efficient algorithm for solving the problem with location constraints.

cs.IT↗
arXiv2024

Guided Real Image Dehazing using YCbCr Color Space

Image dehazing, particularly with learning-based methods, has gained significant attention due to its importance in real-world applications. However, relying solely on the RGB color space often fall short, frequently leaving residual haze. This arises from two main issues: the difficulty in obtaining clear textural features from hazy RGB images and the complexity of acquiring real haze/clean image pairs outside controlled environments like smoke-filled scenes. To address these issues, we first propose a novel Structure Guided Dehazing Network (SGDN) that leverages the superior structural properties of YCbCr features over RGB. It comprises two key modules: Bi-Color Guidance Bridge (BGB) and Color Enhancement Module (CEM). BGB integrates a phase integration module and an interactive attention module, utilizing the rich texture features of the YCbCr space to guide the RGB space, thereby recovering clearer features in both frequency and spatial domains. To maintain tonal consistency, CEM further enhances the color perception of RGB features by aggregating YCbCr channel information. Furthermore, for effective supervised learning, we introduce a Real-World Well-Aligned Haze (RW$^2$AH) dataset, which includes a diverse range of scenes from various geographical regions and climate conditions. Experimental results demonstrate that our method surpasses existing state-of-the-art methods across multiple real-world smoke/haze datasets. Code and Dataset: \textcolor{blue}{\url{https://github.com/fiwy0527/AAAI25_SGDN.}}

cs.CV↗
arXiv2024

Lensless speckle reconstructive spectrometer via physics-aware neural network

The speckle field yielded by disordered media is extensively employed for spectral measurements. Existing speckle reconstructive spectrometers (RSs) implemented by neural networks primarily rely on supervised learning, which necessitates large-scale spectra-speckle pairs. However, beyond system stability requirements for prolonged data collection, generating diverse spectra with high resolution and finely labeling them is particularly difficult. A lack of variety in datasets hinders the generalization of neural networks to new spectrum types. Here we avoid this limitation by introducing PhyspeNet, an untrained spectrum reconstruction framework combining a convolutional neural network (CNN) with a physical model of a chaotic optical cavity. Without pre-training and prior knowledge about the spectrum under test, PhyspeNet requires only a single captured speckle for various multi-wavelength reconstruction tasks. Experimentally, we demonstrate a lens-free, snapshot RS system by leveraging the one-to-many mapping between spatial and spectrum domains in a random medium. Dual-wavelength peaks separated by 2 pm can be distinguished, and a maximum working bandwidth of 40 nm is achieved with high measurement accuracy. This approach establishes a new paradigm for neural network-based RS systems, entirely eliminating reliance on datasets while ensuring that computational results exhibit a high degree of generalizability and physical explainability.

physics.optics↗
arXiv2024

Learning Generalized Residual Exchange-Correlation-Uncertain Functional for Density Functional Theory

Density Functional Theory (DFT) stands as a widely used and efficient approach for addressing the many-electron Schrödinger equation across various domains such as physics, chemistry, and biology. However, a core challenge that persists over the long term pertains to refining the exchange-correlation (XC) approximation. This approximation significantly influences the triumphs and shortcomings observed in DFT applications. Nonetheless, a prevalent issue among XC approximations is the presence of systematic errors, stemming from deviations from the mathematical properties of the exact XC functional. For example, although both B3LYP and DM21 (DeepMind 21) exhibit improvements over previous benchmarks, there is still potential for further refinement. In this paper, we propose a strategy for enhancing XC approximations by estimating the neural uncertainty of the XC functional, named Residual XC-Uncertain Functional. Specifically, our approach involves training a neural network to predict both the mean and variance of the XC functional, treating it as a Gaussian distribution. To ensure stability in each sampling point, we construct the mean by combining traditional XC approximations with our neural predictions, mitigating the risk of divergence or vanishing values. It is crucial to highlight that our methodology excels particularly in cases where systematic errors are pronounced. Empirical outcomes from three benchmark tests substantiate the superiority of our approach over existing state-of-the-art methods. Our approach not only surpasses related techniques but also significantly outperforms both the popular B3LYP and the recent DM21 methods, achieving average RMSE improvements of 62\% and 37\%, respectively, across the three benchmarks: W4-17, G21EA, and G21IP.

cs.CE↗
arXiv2024

Multi-View Fusion Neural Network for Traffic Demand Prediction

The extraction of spatial-temporal features is a crucial research in transportation studies, and current studies typically use a unified temporal modeling mechanism and fixed spatial graph for this purpose. However, the fixed spatial graph restricts the extraction of spatial features for similar but not directly connected nodes, while the unified temporal modeling mechanism overlooks the heterogeneity of temporal variation of different nodes. To address these challenges, a multi-view fusion neural network (MVFN) approach is proposed. In this approach, spatial local features are extracted through the use of a graph convolutional network (GCN), and spatial global features are extracted using a cosine re-weighting linear attention mechanism (CLA). The GCN and CLA are combined to create a graph-cosine module (GCM) for the extraction of overall spatial features. Additionally, the multi-channel separable temporal convolutional network (MSTCN) makes use of a multi-channel temporal convolutional network (MTCN) at each layer to extract unified temporal features, and a separable temporal convolutional network (STCN) to extract independent temporal features. Finally, the spatial-temporal feature data is input into the prediction layer to obtain the final result. The model has been validated on two traffic demand datasets and achieved the best prediction accuracy.

cs.CV↗