Jing Huang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jing Huang has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Jianhua Ma
Southern Medical University
2 shared publications - Xiaoli Li
School of Automation Science and Engineering, South China University of Technology, Guangzhou 510641, China
2 shared publications - Zhaoying Bian
Southern Medical University
2 shared publications - Chen Guo
School of Electronics and Information Engineering, Jinggangshan University, No. 28 Xueyuan Road, Qingyuan District, Ji’an 343009, China
1 shared publication - Chenguang Zhao
Beijing Institute for Brain Research, Chinese Academy of Medical Sciences & Peking Union Medical College; Chinese Institute for Brain Research, Beijing
1 shared publication - Dongwei Li
Department of Psychology, Faculty of Arts and Sciences, Beijing Normal University, Zhuhai, China
1 shared publication
Latest Publications
- Open Access
- 13 November 2026
- Published by MDPI AG in Behavioral Sciences
- Vol. 16 (11) , 2155
Information-Based Multivariate Decoding Reveals Neural Encoding of Attentional Interference During Visual Selective Attention
Visual attention is one of the fundamental abilities underlying human perception, learning, and goal-directed behavior, enabling individuals to prioritize information relevant to current goals while effectively suppressing irrelevant or competing stimuli. While previous studies have revealed the role of top-down attentional control mechanisms in visual processing, particularly in the anticipatory activation of brain regions during target expectation, the precise influence of this regulation on the neural decoding of targets and distractors remains underexplored. This study combines a classical visual search task with a developed multivariate decoding framework to systematically investigate the effects of distractor interference on the neural decoding of target location under conditions of prior spatial knowledge. Through multivariate decoding analysis, we compared the decoding accuracy of target locations in both T-only (only one unique shape-defined target) and T-with-D (shape-defined target with a distractor sharing the same shape) conditions. We found that distractor interference significantly reduced the precision of target location decoding and diminished the temporal generalization of target representations, indicating that interference disrupted the stability of neural decoding. Furthermore, decoding accuracy in both conditions was positively correlated with the amplitude of the N2pc component, further demonstrating the critical role of spatially selective attention in target location decoding. Although preprocessing was applied to the existing decoding methods in this study, it enhanced both the stability and decoding accuracy without affecting the overall trend of the results. Overall, these findings suggest that information-based multivariate decoding offers important empirical support for understanding the neural mechanisms of internal spatial attention under distraction.
- 13 November 2026
- Published by MDPI AG in Behavioral Sciences
- Vol. 16 (11) , 2155
Open AccessInformation-Based Multivariate Decoding Reveals Neural Encoding of Attentional Interference During Visual Selective Attention
Visual attention is one of the fundamental abilities underlying human perception, learning, and goal-directed behavior, enabling individuals to prioritize information relevant to current goals while effectively suppressing irrelevant or competing stimuli. While previous studies have revealed the role of top-down attentional control mechanisms in visual processing, particularly in the anticipatory activation of brain regions during target expectation, the precise influence of this regulation on the neural decoding of targets and distractors remains underexplored. This study combines a classical visual search task with a developed multivariate decoding framework to systematically investigate the effects of distractor interference on the neural decoding of target location under conditions of prior spatial knowledge. Through multivariate decoding analysis, we compared the decoding accuracy of target locations in both T-only (only one unique shape-defined target) and T-with-D (shape-defined target with a distractor sharing the same shape) conditions. We found that distractor interference significantly reduced the precision of target location decoding and diminished the temporal generalization of target representations, indicating that interference disrupted the stability of neural decoding. Furthermore, decoding accuracy in both conditions was positively correlated with the amplitude of the N2pc component, further demonstrating the critical role of spatially selective attention in target location decoding. Although preprocessing was applied to the existing decoding methods in this study, it enhanced both the stability and decoding accuracy without affecting the overall trend of the results. Overall, these findings suggest that information-based multivariate decoding offers important empirical support for understanding the neural mechanisms of internal spatial attention under distraction.
- Open Access
- 22 December 2025
- Published by MDPI AG in Remote Sensing
- Vol. 18 (1) , 22
Online Prototype Angular Balanced Self-Distillation for Non-Ideal Annotation in Remote Sensing Image Segmentation
This paper proposes an Online Prototype Angular Balanced Self-Distillation (OPAB) framework to address the challenges posed by non-ideal annotation in remote sensing image semantic segmentation. “Non-ideal annotation” typically refers to scenarios where long-tailed class distributions and label noise coexist in both training and testing sets. Existing methods often tackle these two issues separately, overlooking the conflict between noisy samples and minority classes as well as the unreliable early stopping caused by non-clean validation sets, which exacerbates the model’s tendency to memorize noisy samples. OPAB mitigates the imbalance problem by employing an improved bilateral-branch network (BBN) that integrates max-min angular regularization (MMA) and category-level inverse weighting to achieve balanced hyperspherical representations. The balanced hyperspherical representations further facilitate noise-clean sample separation and early stopping estimation based on large category-wise Local Intrinsic Dimensionality (LID). Moreover, OPAB introduces a bootstrap teacher label refinement strategy coupled with a student full-parameter retraining mechanism to avoid memorizing noisy samples. Experimental results on ISPRS datasets demonstrate that OPAB achieves a 2.0% mIoU improvement under non-ideal annotation conditions and achieves 89% mIoU after cross-set correction, showcasing strong robustness across different backbones and effective iterative calibration capability.
- 22 December 2025
- Published by MDPI AG in Remote Sensing
- Vol. 18 (1) , 22
Open AccessOnline Prototype Angular Balanced Self-Distillation for Non-Ideal Annotation in Remote Sensing Image Segmentation
This paper proposes an Online Prototype Angular Balanced Self-Distillation (OPAB) framework to address the challenges posed by non-ideal annotation in remote sensing image semantic segmentation. “Non-ideal annotation” typically refers to scenarios where long-tailed class distributions and label noise coexist in both training and testing sets. Existing methods often tackle these two issues separately, overlooking the conflict between noisy samples and minority classes as well as the unreliable early stopping caused by non-clean validation sets, which exacerbates the model’s tendency to memorize noisy samples. OPAB mitigates the imbalance problem by employing an improved bilateral-branch network (BBN) that integrates max-min angular regularization (MMA) and category-level inverse weighting to achieve balanced hyperspherical representations. The balanced hyperspherical representations further facilitate noise-clean sample separation and early stopping estimation based on large category-wise Local Intrinsic Dimensionality (LID). Moreover, OPAB introduces a bootstrap teacher label refinement strategy coupled with a student full-parameter retraining mechanism to avoid memorizing noisy samples. Experimental results on ISPRS datasets demonstrate that OPAB achieves a 2.0% mIoU improvement under non-ideal annotation conditions and achieves 89% mIoU after cross-set correction, showcasing strong robustness across different backbones and effective iterative calibration capability.
- Open Access
- 11 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-5
Co-occurrence objects improve visual search precision in real-world scene through spatial memory
Prior research has demonstrated that scene cues within real-world environments significantly expedite visual search tasks, particularly when co-occurrence objects—those that tend to appear concurrently and in close proximity—are present. The underlying mechanisms driving these enhancements remain a subject of ongoing debate. In this study, we recruited forty-five participants to perform a visual search task using real-world scene images, with conditions involving either the presence or absence of co-occurrence objects, while recording their right eye movements. Our findings revealed that the presence of co-occurrence objects exclusively enhanced search accuracy, without impacting response times or fixation durations. As participants engaged in repeated searches, their performance improved, characterized by increased accuracy, reduced response times, fewer fixations, and shorter fixation durations, applicable to both target objects and co-occurrence objects alike. Notably, the effect of co-occurrence objects remained consistent across search iterations, indicating no change with accumulated search experience. Collectively, these outcomes suggest a strategic role for co-occurrence objects in guiding visual search processes, potentially mediated through the spatial processing of target objects. This study contributes to our understanding of how contextual cues influence visual attention and search precision in complex environments.
- 11 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-5
Open AccessCo-occurrence objects improve visual search precision in real-world scene through spatial memory
Prior research has demonstrated that scene cues within real-world environments significantly expedite visual search tasks, particularly when co-occurrence objects—those that tend to appear concurrently and in close proximity—are present. The underlying mechanisms driving these enhancements remain a subject of ongoing debate. In this study, we recruited forty-five participants to perform a visual search task using real-world scene images, with conditions involving either the presence or absence of co-occurrence objects, while recording their right eye movements. Our findings revealed that the presence of co-occurrence objects exclusively enhanced search accuracy, without impacting response times or fixation durations. As participants engaged in repeated searches, their performance improved, characterized by increased accuracy, reduced response times, fewer fixations, and shorter fixation durations, applicable to both target objects and co-occurrence objects alike. Notably, the effect of co-occurrence objects remained consistent across search iterations, indicating no change with accumulated search experience. Collectively, these outcomes suggest a strategic role for co-occurrence objects in guiding visual search processes, potentially mediated through the spatial processing of target objects. This study contributes to our understanding of how contextual cues influence visual attention and search precision in complex environments.
- Open Access
- 8 March 2023
- Published by Public Library of Science (PLoS) in PLOS Biology
- Vol. 21 (3) , e3002014
Suppression of distracting inputs by visual-spatial cues is driven by anticipatory alpha activity
A growing body of research demonstrates that distracting inputs can be proactively suppressed via spatial cues, nonspatial cues, or experience, which are governed by more than one top-down mechanism of attention. However, how the neural mechanisms underlying spatial distractor cues guide proactive suppression of distracting inputs remains unresolved. Here, we recorded electroencephalography signals from 110 participants in 3 experiments to identify the role of alpha activity in proactive distractor suppression induced by spatial cues and its influence on subsequent distractor inhibition. Behaviorally, we found novel changes in the spatial proximity of the distractor: Cueing distractors far away from the target improves search performance for the target, while cueing distractors close to the target hampers performance. Crucially, we found dynamic characteristics of spatial representation for distractor suppression during anticipation. This result was further verified by alpha power increased relatively contralateral to the cued distractor. At both the between- and within-subjects levels, we found that these activities further predicted the decrement of the subsequent PDcomponent, which was indicative of reduced distractor interference. Moreover, anticipatory alpha activity and its link with the subsequent PDcomponent were specific to the high predictive validity of distractor cue. Together, our results reveal the underlying neural mechanisms by which cueing the spatial distractor may contribute to reduced distractor interference. These results also provide evidence supporting the role of alpha activity as gating by proactive suppression.
- 8 March 2023
- Published by Public Library of Science (PLoS) in PLOS Biology
- Vol. 21 (3) , e3002014
Open AccessSuppression of distracting inputs by visual-spatial cues is driven by anticipatory alpha activity
A growing body of research demonstrates that distracting inputs can be proactively suppressed via spatial cues, nonspatial cues, or experience, which are governed by more than one top-down mechanism of attention. However, how the neural mechanisms underlying spatial distractor cues guide proactive suppression of distracting inputs remains unresolved. Here, we recorded electroencephalography signals from 110 participants in 3 experiments to identify the role of alpha activity in proactive distractor suppression induced by spatial cues and its influence on subsequent distractor inhibition. Behaviorally, we found novel changes in the spatial proximity of the distractor: Cueing distractors far away from the target improves search performance for the target, while cueing distractors close to the target hampers performance. Crucially, we found dynamic characteristics of spatial representation for distractor suppression during anticipation. This result was further verified by alpha power increased relatively contralateral to the cued distractor. At both the between- and within-subjects levels, we found that these activities further predicted the decrement of the subsequent PDcomponent, which was indicative of reduced distractor interference. Moreover, anticipatory alpha activity and its link with the subsequent PDcomponent were specific to the high predictive validity of distractor cue. Together, our results reveal the underlying neural mechanisms by which cueing the spatial distractor may contribute to reduced distractor interference. These results also provide evidence supporting the role of alpha activity as gating by proactive suppression.
- Open Access
- 23 October 2015
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 10 (10) , e0140579
An Efficient Augmented Lagrangian Method for Statistical X-Ray CT Image Reconstruction
Statistical iterative reconstruction (SIR) for X-ray computed tomography (CT) under the penalized weighted least-squares criteria can yield significant gains over conventional analytical reconstruction from the noisy measurement. However, due to the nonlinear expression of the objective function, most exiting algorithms related to the SIR unavoidably suffer from heavy computation load and slow convergence rate, especially when an edge-preserving or sparsity-based penalty or regularization is incorporated. In this work, to address abovementioned issues of the general algorithms related to the SIR, we propose an adaptive nonmonotone alternating direction algorithm in the framework of augmented Lagrangian multiplier method, which is termed as “ALM-ANAD”. The algorithm effectively combines an alternating direction technique with an adaptive nonmonotone line search to minimize the augmented Lagrangian function at each iteration. To evaluate the present ALM-ANAD algorithm, both qualitative and quantitative studies were conducted by using digital and physical phantoms. Experimental results show that the present ALM-ANAD algorithm can achieve noticeable gains over the classical nonlinear conjugate gradient algorithm and state-of-the-art split Bregman algorithm in terms of noise reduction, contrast-to-noise ratio, convergence rate, and universal quality index metrics.
- 23 October 2015
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 10 (10) , e0140579
Open AccessAn Efficient Augmented Lagrangian Method for Statistical X-Ray CT Image Reconstruction
Statistical iterative reconstruction (SIR) for X-ray computed tomography (CT) under the penalized weighted least-squares criteria can yield significant gains over conventional analytical reconstruction from the noisy measurement. However, due to the nonlinear expression of the objective function, most exiting algorithms related to the SIR unavoidably suffer from heavy computation load and slow convergence rate, especially when an edge-preserving or sparsity-based penalty or regularization is incorporated. In this work, to address abovementioned issues of the general algorithms related to the SIR, we propose an adaptive nonmonotone alternating direction algorithm in the framework of augmented Lagrangian multiplier method, which is termed as “ALM-ANAD”. The algorithm effectively combines an alternating direction technique with an adaptive nonmonotone line search to minimize the augmented Lagrangian function at each iteration. To evaluate the present ALM-ANAD algorithm, both qualitative and quantitative studies were conducted by using digital and physical phantoms. Experimental results show that the present ALM-ANAD algorithm can achieve noticeable gains over the classical nonlinear conjugate gradient algorithm and state-of-the-art split Bregman algorithm in terms of noise reduction, contrast-to-noise ratio, convergence rate, and universal quality index metrics.
- 1 January 2013
- Published by Springer Nature
- p. 838-841
Perfusion Computed Tomography Image Reconstruction Using Spatio-Temporal Constrained Edge-preserving Prior
- 1 January 2013
- Published by Springer Nature
- p. 838-841
Perfusion Computed Tomography Image Reconstruction Using Spatio-Temporal Constrained Edge-preserving Prior