RZRui Zhang

Rui Zhang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Rui Zhang has a total of 73 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • A Reliability-Guided Fusion and Dynamic Dual-Domain Mask-Modulated Shearlet Network for Medical Image Segmentation

    by Qingting Jiang, Hailiang Ye, Rui Zhang, Feilong Cao

    Deep learning methods are becoming increasingly important for medical image segmentation. However, previous approaches often performed poorly in spatial-frequency-domain fusion and feature fusion, leading to inaccurate lesion localization, hazy boundaries, and low accuracy. This paper proposes a reliability-guided fusion and dynamic dual-domain mask-modulated Shearlet network for medical image segmentation (RF-DMSNet for short). Its two primary components are the reliability-guided multi-scale feature fusion module (RGMF) and the dynamic dual-domain mask with a coordinate attention-modulated Shearlet operator (DAMS). The former module exhibits reliability in both the channel and spatial dimensions, allowing adaptive weighted fusion to increase feature quality. Meanwhile, the latter uses a dynamic dual-domain mask and coordinate attention to improve lesion boundary and location segmentation performance. The frequency-domain re-enhancement module (FREM) and critical feature guided module (CFGM) are designed using DAMS. These two modules work together to optimize a unified, detailed enhancement module (DEM), enabling progressive refinement of segmentation predictions from coarse to fine. Extensive experimental results demonstrate that RF-DMSNet achieves superior segmentation performance compared to state-of-the-art approaches for medical image segmentation.

  • Histamine regulation in shaping spindle refractoriness: a computational modeling study

    by Bo Wang, Qiang Li, Wen-Hua Wang, Wan-Rong Zan, Yi-Ming Li, Jiang-Ling Song, Rui Zhang

  • Perceptual Gating Mamba Network for Salient Object Detection in Remote Sensing Observation

    by Jian Yang, Dawei Song, Qiurong Yan, Kang Liu, Rui Zhang

    Salient object detection in remote sensing images (RSI-SOD) plays a vital role in enabling large-scale environmental perception for embodied intelligence systems. Despite growing interest, existing methods often underperform in complex urban settings characterized by cluttered backgrounds and significant scale variations, suffering from either blurred object boundaries or high computational costs. To address these issues, we propose a perceptual gating mamba network (PGMNet) that accurately and efficiently detects objects with a global receptive field. The network first employs a spatial neighboring scanning strategy to extract structurally coherent image patches. A gated focus-aware mamba fusion module is then designed to enhance spatial coherence across these patches through joint channel focusing, spatial attention, and a gating mechanism, thereby improving localization accuracy. Furthermore, a structure-preserving upsampling aggregation module is developed to reconstruct complete object contours by aligning and aggregating multi-level features during decoding. Extensive experiments on two benchmark datasets demonstrate that PGMNet outperforms 12 competing methods, validating its effectiveness and robustness in complex urban environments.


  • Ripple-Spikes Localize the Epileptogenic Zone and Its Interpretation: Integrating Data-Driven and Model-Driven Methods

    by Qiang Li, Min Pan, Bo Wang, Wenhua Wang, Jiangling Song, Rui Zhang

    Objective: First, this study evaluates whether ripples co-occurring with epileptiform spikes (i.e., ripple-spikes) in scalp EEG provide a reliable biomarker for epileptogenic zone (EZ) localization compared to ripples or spikes alone. Then, this study develops a novel computational hippocampo-cortical model to explore the mechanisms underlying ripple-spikes and provides a mechanistic explanation for why ripple-spikes locate EZ more effectively than ripples or spikes. Methods: We apply automated detectors to identify ripples, spikes, and ripple-spikes on scalp EEG recordings in patients with focal epilepsy from two public datasets, and further perform a statistical analysis of the rates of these three events in EZ and non-epileptogenic zone (NEZ); the model consists of a hippocampal CA1 module and a cortical module, where an electrical synaptic current is introduced in the CA1 module and the reducing of GABAergic inhibition is modeled in the cortical module. Results: Real data analysis results show a significantly higher ratio of ripple-spike rate in EZ to that in NEZ, compared to ripple ($p=0.01$) or spike ($p=0.008$); simulation results demonstrate that: 1) ripple-spikes can be mediated by the hippocampo-cortical connection strength; 2) and such connection may be stronger for EZ than that for NEZ, leading to a higher ratio of ripple-spike rate. Conclusion: Ripple-spikes in scalp EEG may serve as an improved biomarker for EZ identification, potentially explained by stronger hippocampo-cortical connection for EZ. Significance: Providing a meaningful attempt in terms of the computational method to study ripple-spikes, offering both data and theoretical support for advancing noninvasive methods of EZ localization.

  • 跨尺度时序交互与边界概率细化的连续动作分割与预测

    by Hu Zhonghua, 中华 西北大学大数据与认知计算研究中心,陕西西安710127ResearchCenterforBigDataandCognitiveComputing,NorthwestUniversity,Xi’an,Shaanxi710127,China, Li Xiaoning, Song Jiangling, Zhang Rui

    目的针对连续视频动作分割与预测任务中局部动作动态与长时语义上下文之间交互建模不足、连续动作过渡阶段边界表征不足以及未来序列预测稳定性不足等问题,提出一种融合跨尺度时序交互与边界概率细化的连续动作分割与预测方法。方法面向动作分割与预测两类任务,构建由局部时空编码、跨尺度时序交互和边界概率细化组成的统一框架。对于分割任务,采用膨胀三维卷积网络(inflated 3D convolutional network,I3D)对RGB视频序列进行时空特征编码;对于预测任务,采用X3D网络对RGB序列与光流序列同时进行时空特征提取,并通过双向交叉注意力机制实现多模态信息融合。在此基础上,设计跨尺度时序交互模块,对不同层级时空特征进行编码—解码式建模,刻画不同时间尺度下的上下文依赖关系,增强模型对局部动态变化与全局时序结构的协同表征能力;进一步引入边界概率细化模块,对动作边界进行显式建模,从而提升动作分割与预测结果的稳定性及边界定位精度。结果在50Salads、GTEA、Breakfast等生活场景连续动作分割数据集以及JIGSAWS手术手势数据集上进行了实验验证。所提方法在50Salads、GTEA、Breakfast和JIGSAWS分割任务上的准确率(accuracy,Acc)分别达到92.3%、88.0%、83.2%和90.5%,Edit分别达到89.0%、95.8%、84.7%和95.2%;在JIGSAWS手术动作预测任务中,预测准确率达到90.3%。实验结果表明,所提方法在多数指标上优于现有代表性方法。结论所提方法能够有效兼顾局部动态特征建模、长程时序依赖建模与边界结构优化,在连续动作分割与预测任务中表现出良好的泛化能力与稳定性。

  • EEG Quantification for Color Vision via Steady-State Visual Evoked Potential

    by Xiaowei Zheng, Guanghua Xu, Rui Zhang

    Human color vision depends on the interaction of L‒, M‒, and S‒cone photoreceptors, and impairments in this system lead to color vision deficiencies (CVDs). Traditional psychophysical tests are subjective and qualitative. This study aimed to develop an objective and quantitative evaluation method for human color vision based on Electroencephalography (EEG). Steady-state visual evoked potentials (SSVEPs) were induced by a series of luminance-color coupled stimuli (i.e., red-green pattern-reversal checkerboards with various red-green luminance ratios) for ten dyschromatopsia and seven healthy controls. Then, based on the theory of equiluminance point, this study established a luminance ratio-SSVEP response curve model and proposed the corresponding equiluminance point threshold determination criteria. After obtaining the equiluminance value, a color vision index (CVI) was designed to characterize the functional state of color vision objectively. Meanwhile, the psychophysical detection means of Farnsworth–Munsell (FM) 100-hue testing was carried out to obtain subjective results. The results of the equiluminance value show a difference among normal, deutan, and protan (P < 0.001, respectively), revealing that equiluminance points can qualitatively distinguish normal and abnormal color vision as well as determine the abnormal type. The CVIs show strong correlations with psychophysical test results of both total error scores (TES) and confusion index (C-index) from the FM 100-hue testing, with Pearson’s r values of up to 0.938, revealing that this proposed CVI can be regarded as a quantitative indicator for assessing color vision. The findings of this study indicated that SSVEP might be a potential biomarker to measure human color vision function objectively and quantitatively.


  • Enhancing performance of SSVEP-based visual acuity via exploiting subject-specific information

    by Xiaowei Zheng, Boyu Wen, Xin Yan, Guanghua Xu, Rui Zhang

    This study sought to enhance visual acuity assessment using steady-state visual evoked potentials (SSVEPs) through subject-specific training methods. SSVEPs were elicited from eleven subjects using the vertical sinusoidal gratings at six various spatial frequency steps, and then the classical approach of Oz single-channel, the spatial filtering method of canonical correlation analysis (CCA), and five subject-specific training methods, i.e., individual template-based canonical correlation analysis (IT-CCA), multi-way canonical correlation analysis (MwayCCA), multi-set canonical correlation analysis (MsetCCA), task-related component analysis (TRCA), and correlated component analysis (CORCA), were used as preprocessed methods for six-channel SSVEP signals. Subsequently, by comparing the SSVEP response characteristics, MwayCCA and TRCA were selected for further processing with Oz-channel and CCA as the controls. After carrying out the SSVEP visual acuity estimation criterion, Bland-Altman analysis showed an agreement of 0.201, 0.195, 0.188, and 0.196 logMAR between the subjective Freiburg Visual Acuity and Contrast Test (FrACT) and the objective SSVEP visual acuity for Oz-channel, CCA, MwayCCA, and TRCA, respectively, demonstrating that the subject-specific training method of MwayCCA showed the optimal performance in SSVEP-based visual acuity assessment. This study demonstrated that subject-specific training methods enhance SSVEP-based visual acuity assessment and recommended MwayCCA as the preferred approach for signal preprocessing in such assessments.

  • Bidirectional Multiscale Efficient Dilated Convolutional Recurrent Neural Network Improved by Swarm Intelligence Optimization

    by Qinwei Fan, Shuai Zhao, Jacek M. Zurada, Tingwen Huang, Xiaolong Qin, Rui Zhang

    In recent years, bidirectional convolutional recurrent neural networks (RNNs) have made significant breakthroughs in addressing a wide range of challenging problems related to time series and prediction applications. However, the performance of the models is highly dependent on the hyperparameters chosen. Hence, we propose an automatic method for hyperparameter optimization and apply a bidirectional convolutional RNN based on the improved swarm intelligence optimization (sparrow search) to solve regression prediction problems. Specifically, a parallel multiscale dilated convolution (PMDC) module was designed to capture both local and global spatial correlations. This method utilizes convolution with different dilation rates to expand the receptive field without increasing the complexity of the model. Meanwhile, it integrates parallel multiscale structures to extract features at different scales and enhance the model’s understanding of the input data. Then, the bidirectional gated recurrent units (BGRUs) learn temporal information from the convolutional features. To address the limitations of empirical hyperparameter selection, such as slow training and low efficiency, a novel PMDC-BGRU model integrated with a pretrained sparrow search algorithm (SSA) was proposed for hyperparameter optimization. Finally, experiments on multiple datasets verified the superiority of the algorithm and explained the flexibility of intelligent optimization algorithms in solving model parameter optimization.


  • Understanding the Spatio-Temporal Coupling of Spikes and Spindles in Focal Epilepsy Through a Network-Level Computational Model

    by Min Pan, Qiang Li, Jiangling Song, Bo Wang, Wenhua Wang, Rui Zhang

    The electrophysiological findings have shown that epileptiform spikes triggering sleep spindles within 1[ ]s across multiple channels are commonly observed during sleep in focal epilepsy (FE). Such spatio-temporal couplings of spikes and spindles (STCSSs) are defined as a kind of pathological waves, and frequent emergence of them may cause the degradation of cognitive function for FE patients. However, the neural mechanisms underlying STCSSs are not well understood. To this end, this work first develops a neural mass network model for focal epilepsy (FE-NMNM) with multiple thalamocortical columns being its nodes and the long-range synaptic interactions of thalamocortical columns being its edges, where each thalamocortical column is extended on the basis of Costa model and then they are connected through excitatory synapses between pyramidal cells. Then, how the cortico-cortical connectivity affects the evolution of STCSSs across the network is especially discussed by simulations in two cases, where the inter-ictal state and the ictal state are considered separately. Simulation results demonstrate that: (1) the more STCSSs occur in a more extensive area when the cortico-cortical connectivity becomes stronger, and the significant increase of coupling discharges is attributed to the presence of abundant spikes; (2) when the connectivity is excessively strong, the cortical hyperexcitability will happen, thereby inducing massive spike discharges which may further inhibit the occurrence of spindles, and hence, resulting in the disappearance of STCSSs. The obtained results provide a mechanistic insight into STCSSs, and suggest that such coupling patterns could reflect widespread network dysfunction in FE, thereby potentially advancing therapeutic strategies for FE.