arXiv subjects

Junfeng Zhang

Publications and source records attributed to Junfeng Zhang.

17 recordsLinked to original sources

arXiv2026

Hamilton-Jacobi Reachability-Based Safe Reinforcement Learning for Emergency Collision Avoidance

Emergency collision avoidance under extreme driving conditions demands safety-critical control that accounts for both obstacle proximity and vehicle dynamic stability over a future time horizon, yet existing methods often rely on instantaneous or local safety evaluations. This paper proposes a safe reinforcement learning framework guided by a Hamilton-Jacobi (HJ) reachability based motion safety set that provides forward-looking safety supervision for constrained policy optimization. Specifically, a unified signed safety function is formulated by combining geometric collision margins and chassis stability limits, and is then extended through reachability analysis into a finite-horizon motion safety set that characterizes whether safety can be maintained under future vehicle state evolution. To enable practical computation, the motion safety set is approximated from offline extreme driving data, mitigating the computational burden of grid-based HJ solvers. The learned motion safety set is then embedded as a continuous safety cost into a constrained Markov decision process, and a PID-Lagrangian policy optimization scheme is employed to adaptively regulate the Lagrange multiplier for safety constraint enforcement. Simulation and real-vehicle experiments on low-adhesion obstacle-avoidance scenarios demonstrate that the proposed method achieves higher goal-reaching rates, produces smoother avoidance maneuvers, and maintains larger unified safety margins than baseline methods.

eess.SY↗

arXiv2026

Simple but not Simpler: A Surface-Sliding Method for Finding the Minimum Distance between Two Ellipsoids

We propose a novel iterative process to establish the minimum separation between two ellipsoids. The method maintains one point on each surface and updates their locations in the theta-phi parametric space. The tension along the connecting segment between the two surface points serves as the guidance for the sliding direction, and the distance between them decreases gradually. The minimum distance is established when the connecting segment becomes perpendicular to the ellipsoid surfaces, at which point the net effect of the segment tension disappears and the surface points no longer move. Demonstration examples are carefully designed, and excellent numerical performance is observed, including accuracy, consistency, stability, and robustness. Furthermore, compared to other existing techniques, this surface-sliding approach has several attractive features, such as clear geometric representation, concise formulation, a simple algorithm, and the potential to be extended straightforwardly to other situations. This method is expected to be useful for future studies in computer graphics, engineering design, material modeling, and scientific simulations.

cs.CG↗

arXiv2025

A Cross-Hierarchical Difference Feature Fusion Network Based on Multiscale Encoder-Decoder for Hyperspectral Change Detection

Hyperspectral change detection (HCD) is one of the core applications of remote sensing images, holding significant research value in fields like environmental monitoring and disaster assessment. However, existing methods often suffer from incomplete capture of multiscale spatial-spectral features and insufficient fusion of differential feature information. To address these challenges, this paper proposes a Cross-Hierarchical Differential Feature Fusion Network (CHDFFN) based on a multiscale encoder-decoder. Firstly, a multiscale feature extraction subnetwork is designed, taking the customized encoder-decoder as the backbone, combined with residual connections and the proposed dual-core channel-spatial attention module to achieve multi-level extraction and initial integration of spatial-spectral features. The encoder embeds convolutional blocks with different receptive field sizes to capture multiscale representations from shallow details to deep semantics. The decoder fuses the encoder's output via skip connections to gradually restore spatial resolution while suppressing background noise and redundancy. To enhance the model's ability to capture differential features between bi-temporal hyperspectral images, a spatial-spectral change feature learning module is designed to learn hierarchical change representations. Additionally, an adaptive high-level feature fusion module is proposed, dynamically balancing the contribution of hierarchical differential features by adaptively assigning weights, which effectively strengthens the model's capability to characterize complex change patterns. Finally, experiments on four public hyperspectral datasets show that compared with some state-of-the-art methods, the average maximum improvements of OA, KC, and F1 are 4.61%, 19.79%, and 18.90% respectively, verifying the model's effectiveness.

cs.CV↗

arXiv2025

New high-dimensional generalizations of Nesbitt's inequality and relative applications

Two kinds of novel generalizations of Nesbitt's inequality are explored in various cases regarding dimensions and parameters in this article. Some other cases are also discussed elaborately by using the semiconcave-semiconvex theorem. The general inequalities are then employed to deduce some alternate inequalities and mathematical competition questions. At last, a relation about Hurwitz-Lerch zeta functions is obtained.

math.GM↗

arXiv2024

One-step Structure Prediction and Screening for Protein-Ligand Complexes using Multi-Task Geometric Deep Learning

Understanding the structure of the protein-ligand complex is crucial to drug development. Existing virtual structure measurement and screening methods are dominated by docking and its derived methods combined with deep learning. However, the sampling and scoring methodology have largely restricted the accuracy and efficiency. Here, we show that these two fundamental tasks can be accurately tackled with a single model, namely LigPose, based on multi-task geometric deep learning. By representing the ligand and the protein pair as a graph, LigPose directly optimizes the three-dimensional structure of the complex, with the learning of binding strength and atomic interactions as auxiliary tasks, enabling its one-step prediction ability without docking tools. Extensive experiments show LigPose achieved state-of-the-art performance on major tasks in drug research. Its considerable improvements indicate a promising paradigm of AI-based pipeline for drug development.

cs.AI↗

arXiv2022

Transformers in Medical Image Analysis: A Review

Transformers have dominated the field of natural language processing, and recently impacted the computer vision area. In the field of medical image analysis, Transformers have also been successfully applied to full-stack clinical applications, including image synthesis/reconstruction, registration, segmentation, detection, and diagnosis. Our paper aims to promote awareness and application of Transformers in the field of medical image analysis. Specifically, we first overview the core concepts of the attention mechanism built into Transformers and other basic components. Second, we review various Transformer architectures tailored for medical image applications and discuss their limitations. Within this review, we investigate key challenges revolving around the use of Transformers in different learning paradigms, improving the model efficiency, and their coupling with other techniques. We hope this review can give a comprehensive picture of Transformers to the readers in the field of medical image analysis.

cs.CV↗

arXiv2021

Comb-based photonic neural population for parallel and nonlinear processing

It is believed that neural information representation and processing relies on the neural population instead of a single neuron. In neuromorphic photonics, photonic neurons in the form of nonlinear responses have been extensively studied in single devices and temporal nodes. However, to construct a photonic neural population (PNP), the process of scaling up and massive interconnections remain challenging considering the physical complexity and response latency. Here, we propose a comb-based PNP interconnected by carrier coupling with superior scalability. Two unique properties of neural population are theoretically and experimentally demonstrated in the comb-based PNP, including nonlinear response curves and population activities coding. A classification task of three input patterns with dual radio-frequency (RF) tones is successfully implemented in a real-time manner, which manifests the comb-based PNP can make effective use of the ultra-broad bandwidth of photonics for parallel and nonlinear processing.

cs.ET↗

arXiv2021

Cross-Modality Brain Tumor Segmentation via Bidirectional Global-to-Local Unsupervised Domain Adaptation

Accurate segmentation of brain tumors from multi-modal Magnetic Resonance (MR) images is essential in brain tumor diagnosis and treatment. However, due to the existence of domain shifts among different modalities, the performance of networks decreases dramatically when training on one modality and performing on another, e.g., train on T1 image while performing on T2 image, which is often required in clinical applications. This also prohibits a network from being trained on labeled data and then transferred to unlabeled data from a different domain. To overcome this, unsupervised domain adaptation (UDA) methods provide effective solutions to alleviate the domain shift between labeled source data and unlabeled target data. In this paper, we propose a novel Bidirectional Global-to-Local (BiGL) adaptation framework under a UDA scheme. Specifically, a bidirectional image synthesis and segmentation module is proposed to segment the brain tumor using the intermediate data distributions generated for the two domains, which includes an image-to-image translator and a shared-weighted segmentation network. Further, a global-to-local consistency learning module is proposed to build robust representation alignments in an integrated way. Extensive experiments on a multi-modal brain MR benchmark dataset demonstrate that the proposed method outperforms several state-of-the-art unsupervised domain adaptation methods by a large margin, while a comprehensive ablation study validates the effectiveness of each key component. The implementation code of our method will be released at \url{https://github.com/KeleiHe/BiGL}.

cs.CV↗

arXiv2021

MetricUNet: Synergistic Image- and Voxel-Level Learning for Precise CT Prostate Segmentation via Online Sampling

Fully convolutional networks (FCNs), including UNet and VNet, are widely-used network architectures for semantic segmentation in recent studies. However, conventional FCN is typically trained by the cross-entropy or Dice loss, which only calculates the error between predictions and ground-truth labels for pixels individually. This often results in non-smooth neighborhoods in the predicted segmentation. To address this problem, we propose a two-stage framework, with the first stage to quickly localize the prostate region and the second stage to precisely segment the prostate by a multi-task UNet architecture. We introduce a novel online metric learning module through voxel-wise sampling in the multi-task network. Therefore, the proposed network has a dual-branch architecture that tackles two tasks: 1) a segmentation sub-network aiming to generate the prostate segmentation, and 2) a voxel-metric learning sub-network aiming to improve the quality of the learned feature space supervised by a metric loss. Specifically, the voxel-metric learning sub-network samples tuples (including triplets and pairs) in voxel-level through the intermediate feature maps. Unlike conventional deep metric learning methods that generate triplets or pairs in image-level before the training phase, our proposed voxel-wise tuples are sampled in an online manner and operated in an end-to-end fashion via multi-task learning. To evaluate the proposed method, we implement extensive experiments on a real CT image dataset consisting of 339 patients. The ablation studies show that our method can effectively learn more representative voxel-level features compared with the conventional learning methods with cross-entropy or Dice loss. And the comparisons show that the proposed method outperforms the state-of-the-art methods by a reasonable margin.

eess.IV↗

arXiv2020

Detection of Intermittent Faults Based on an Optimally Weighted Moving Average T^2 Control Chart with Stationary Observations

The moving average (MA)-type scheme, also known as the smoothing method, has been well established within the multivariate statistical process monitoring (MSPM) framework since the 1990s. However, its theoretical basis is still limited to smoothing independent data, and the optimality of its equally or exponentially weighted scheme remains unproven. This paper aims to weaken the independence assumption in the existing MA method, and then extend it to a broader area of dealing with autocorrelated weakly stationary processes. With the discovery of the non-optimality of the equally and exponentially weighted schemes used for fault detection when data have autocorrelation, the essence that they do not effectively utilize the correlation information of samples is revealed, giving birth to an optimally weighted moving average (OWMA) theory. The OWMA method is combined with the Hotelling's $T^2$ statistic to form an OWMA $T^2$ control chart (OWMA-TCC), in order to detect a more challenging type of fault, i.e., intermittent fault (IF). Different from the MA scheme that puts an equal weight on samples within a time window, OWMA-TCC uses correlation (autocorrelation and cross-correlation) information to find an optimal weight vector (OWV) for the purpose of IF detection (IFD). In order to achieve a best IFD performance, the concept of IF detectability is defined and corresponding detectability conditions are provided, which further serve as selection criteria of the OWV. Then, the OWV is given in the form of a solution to nonlinear equations, whose existence is proven with the aid of the Brouwer fixed-point theory. Moreover, symmetrical structure of the OWV is revealed, and the optimality of the MA scheme for any IF directions when data exhibit no autocorrelation is proven.

cs.IT↗

arXiv2020

Detection and Detectability of Intermittent Faults Based on Moving Average T2 Control Charts with Multiple Window Lengths

So far, problems of intermittent fault (IF) detection and detectability have not been fully investigated in the multivariate statistics framework. The characteristics of IFs are small magnitudes and short durations, and consequently traditional multivariate statistical methods using only a single observation are no longer effective. Thus in this paper, moving average T^2 control charts (MA-TCCs) with multiple window lengths, which simultaneously employ a bank of MA-TCCs with different window lengths, are proposed to address the IF detection problem. Methods to reduce false/missing alarms and infer the IFs' appearing and disappearing time instances are presented. In order to analyze the detection capability for IFs, definitions of guaranteed detectability are introduced, which is an extension and generalization of the original fault detectability concept focused on permanent faults (PFs). Then, necessary and sufficient conditions are derived for the detectability of IFs, which may appear and disappear several times with different magnitudes and durations. Based on these conditions, some optimal properties of two important window lengths are further discussed. In this way, a theoretical framework for the analysis of IFs' detectability is established as well as extended discussions on how the theoretical results can be adapted to real-world applications. Finally, simulation studies on a numerical example and the continuous stirred tank reactor (CSTR) process are carried out to show the effectiveness of the developed methods.

eess.SY↗

arXiv2020

Synergistic Learning of Lung Lobe Segmentation and Hierarchical Multi-Instance Classification for Automated Severity Assessment of COVID-19 in CT Images

Understanding chest CT imaging of the coronavirus disease 2019 (COVID-19) will help detect infections early and assess the disease progression. Especially, automated severity assessment of COVID-19 in CT images plays an essential role in identifying cases that are in great need of intensive clinical care. However, it is often challenging to accurately assess the severity of this disease in CT images, due to variable infection regions in the lungs, similar imaging biomarkers, and large inter-case variations. To this end, we propose a synergistic learning framework for automated severity assessment of COVID-19 in 3D CT images, by jointly performing lung lobe segmentation and multi-instance classification. Considering that only a few infection regions in a CT image are related to the severity assessment, we first represent each input image by a bag that contains a set of 2D image patches (with each cropped from a specific slice). A multi-task multi-instance deep network (called M$^2$UNet) is then developed to assess the severity of COVID-19 patients and also segment the lung lobe simultaneously. Our M$^2$UNet consists of a patch-level encoder, a segmentation sub-network for lung lobe segmentation, and a classification sub-network for severity assessment (with a unique hierarchical multi-instance learning strategy). Here, the context information provided by segmentation can be implicitly employed to improve the performance of severity assessment. Extensive experiments were performed on a real COVID-19 CT image dataset consisting of 666 chest CT images, with results suggesting the effectiveness of our proposed method compared to several state-of-the-art methods.

eess.IV↗

arXiv2020

HF-UNet: Learning Hierarchically Inter-Task Relevance in Multi-Task U-Net for Accurate Prostate Segmentation

Accurate segmentation of the prostate is a key step in external beam radiation therapy treatments. In this paper, we tackle the challenging task of prostate segmentation in CT images by a two-stage network with 1) the first stage to fast localize, and 2) the second stage to accurately segment the prostate. To precisely segment the prostate in the second stage, we formulate prostate segmentation into a multi-task learning framework, which includes a main task to segment the prostate, and an auxiliary task to delineate the prostate boundary. Here, the second task is applied to provide additional guidance of unclear prostate boundary in CT images. Besides, the conventional multi-task deep networks typically share most of the parameters (i.e., feature representations) across all tasks, which may limit their data fitting ability, as the specificities of different tasks are inevitably ignored. By contrast, we solve them by a hierarchically-fused U-Net structure, namely HF-UNet. The HF-UNet has two complementary branches for two tasks, with the novel proposed attention-based task consistency learning block to communicate at each level between the two decoding branches. Therefore, HF-UNet endows the ability to learn hierarchically the shared representations for different tasks, and preserve the specificities of learned representations for different tasks simultaneously. We did extensive evaluations of the proposed method on a large planning CT image dataset, including images acquired from 339 patients. The experimental results show HF-UNet outperforms the conventional multi-task network architectures and the state-of-the-art methods.

eess.IV↗

arXiv2019

A Physics-Based Estimation of Mean Curvature Normal Vector for Triangulated Surfaces

In this note, we derive an approximation for the mean curvature normal vector on vertices of triangulated surface meshes from the Young-Laplace equation and the force balance principle. We then demonstrate that the approximation expression from our physics-based derivation is equivalent to the discrete Laplace-Beltrami operator approach in the literature. This work, in addition to providing an alternative expression to calculate the mean curvature normal vector, can be further extended to other mesh structures, including non-triangular and heterogeneous meshes.

math.DG↗

arXiv2018

LEARN: Learned Experts' Assessment-based Reconstruction Network for Sparse-data CT

Compressive sensing (CS) has proved effective for tomographic reconstruction from sparsely collected data or under-sampled measurements, which are practically important for few-view CT, tomosynthesis, interior tomography, and so on. To perform sparse-data CT, the iterative reconstruction commonly use regularizers in the CS framework. Currently, how to choose the parameters adaptively for regularization is a major open problem. In this paper, inspired by the idea of machine learning especially deep learning, we unfold a state-of-the-art "fields of experts" based iterative reconstruction scheme up to a number of iterations for data-driven training, construct a Learned Experts' Assessment-based Reconstruction Network ("LEARN") for sparse-data CT, and demonstrate the feasibility and merits of our LEARN network. The experimental results with our proposed LEARN network produces a competitive performance with the well-known Mayo Clinic Low-Dose Challenge Dataset relative to several state-of-the-art methods, in terms of artifact reduction, feature preservation, and computational speed. This is consistent to our insight that because all the regularization terms and parameters used in the iterative reconstruction are now learned from the training data, our LEARN network utilizes application-oriented knowledge more effectively and recovers underlying images more favorably than competing algorithms. Also, the number of layers in the LEARN network is only 12, reducing the computational complexity of typical iterative algorithms by orders of magnitude.

physics.med-ph↗

arXiv2018

Quantum oscillation in carrier transport in two-dimensional junctions

Two-dimensional (2D) device structures have recently attracted considerable attention. Here, we show that most 2D device structures, regardless vertical or lateral, act as a lateral monolayer-bilayer-monolayer junction in their operation. In particular, a vertical structure cannot function as a vertical junction as having been widely believed in the literature. Moreover, due to a larger electrostatic screening, the bilayer region in the junction always has a smaller band gap than its monolayer counterpart. As a result, a potential well, aside from the usual potential barrier, will form universally in the bilayer region to affect the hole or electron quantum transport in the form of transmission or reflection. Taking black phosphorus as an example, we show that an oscillation in the transmission coefficient can be clearly resolved in a two-electrode prototypical device by non-equilibrium Green function combined with density functional theory calculations and the results can be qualitatively understood using a simple quantum well model. The presence of the quantum well is vital to 2D device design, including the effective tuning of quantum transmission by a vertical electric field.

cond-mat.mes-hall↗

arXiv2016

Band alignment of two-dimensional lateral heterostructures

Recent experimental synthesis of two-dimensional (2D) heterostructures opens a door to new opportunities in tailoring the electronic properties for novel 2D devices. Here, we show that a wide range of lateral 2D heterostructures could have a prominent advantage over the traditional three-dimensional (3D) heterostructures, because their band alignments are insensitive to the interfacial conditions. They should be at the Schottky-Mott limits for semiconductor-metal junctions and at the Anderson limits for semiconductor junctions, respectively. This fundamental difference from the 3D heterostructures is rooted in the fact that, in the asymptotic limit of large distance, the effect of the interfacial dipole vanishes for 2D systems. Due to the slow decay of the dipole field and the dependence on the vacuum thickness, however, studies based on first-principles calculations often failed to reach such a conclusion. Taking graphene/hexagonal-BN and MoS2/WS2 lateral heterostructures as the respective prototypes, we show that the converged junction width can be order of magnitude longer than that for 3D junctions. The present results provide vital guidance to high-quality transport devices wherever a lateral 2D heterostructure is involved.

physics.comp-ph↗