Nan Xia
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Nan Xia has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Jiahong Jiang
School of Information Science and Engineering, Dalian Polytechnic University, Dalian, China
3 shared publications - Baohui Xing
School of Information Science and Engineering, Dalian Polytechnic University, Dalian, China
2 shared publications - Bo Li
School of Information Science and Engineering, Dalian Polytechnic University, Dalian 116034 China
2 shared publications - Jue Wang
School of Information Science and Engineering, Dalian Polytechnic University, Dalian, China
2 shared publications Bo Li
School of Information Science and Engineering, Dalian Polytechnic University, Dalian, China 116034
1 shared publicationDanyang Gao
School of Information Science and Engineering, Dalian Polytechnic University, Dalian, China
1 shared publication
Latest Publications
- 3 October 2026
- Published by Elsevier in Applied Soft Computing
- Vol. 204
Multimodal speech emotion recognition based on dual cross-modal attention mechanism
- 3 October 2026
- Published by Elsevier in Applied Soft Computing
- Vol. 204
Multimodal speech emotion recognition based on dual cross-modal attention mechanism
- 22 September 2026
- Published by Elsevier in Neurocomputing
- Vol. 708
DualMamba: Uncertainty-guided pose-velocity state space modeling for 3D human pose estimation
- 22 September 2026
- Published by Elsevier in Neurocomputing
- Vol. 708
DualMamba: Uncertainty-guided pose-velocity state space modeling for 3D human pose estimation
- Open Access
- 20 August 2026
- Published by Springer Nature in Journal of King Saud University Computer and Information Sciences
- Vol. 38 (7) , 718
Importance relation guided spatio-temporal adaptive modeling framework for irregular 2D human pose estimation
- 20 August 2026
- Published by Springer Nature in Journal of King Saud University Computer and Information Sciences
- Vol. 38 (7) , 718
Open AccessImportance relation guided spatio-temporal adaptive modeling framework for irregular 2D human pose estimation
- 20 January 2026
- Published by Elsevier in Measurement
- Vol. 266
Automotive radar co-channel interference mitigation and target detection based on third-order cumulant and neural network
- 20 January 2026
- Published by Elsevier in Measurement
- Vol. 266
Automotive radar co-channel interference mitigation and target detection based on third-order cumulant and neural network
- Open Access
- 5 January 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (1) , 178
A Feature-Enhanced Network-Based Target Detection Method for SAR Images of Ships in Complex Scenes
In the context of ship target detection with Synthetic Aperture Radar (SAR) images, misdetection and missed detection are often caused by complex background interference and the variability in target size. To address these challenges, this paper proposes an innovative method based on image enhancement and feature fusion to reduce background noise and effectively handle the detection confusion caused by differences in ship sizes. Firstly, a feature-aware enhancement network is introduced, which preserves and strengthens the edge information of the target objects. Secondly, during the feature extraction phase, a dynamic hierarchical extraction module is proposed, significantly improving the feature capture ability of convolutional neural networks and overcoming the limitations of traditional fixed kernel receptive fields. Finally, a feature fusion module based on attention gating is employed to fully leverage the complementary information between the original and enhanced images, achieving precise modeling and efficient fusion of inter-feature correlations. The proposed method is integrated with the YOLOv8 detection framework for target detection. Experimental results in the publicly available SSDD and HRSID datasets demonstrate detection accuracies of 97.9% and 93.2%, respectively, thus validating the superiority and robustness of the proposed method.
- 5 January 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (1) , 178
Open AccessA Feature-Enhanced Network-Based Target Detection Method for SAR Images of Ships in Complex Scenes
In the context of ship target detection with Synthetic Aperture Radar (SAR) images, misdetection and missed detection are often caused by complex background interference and the variability in target size. To address these challenges, this paper proposes an innovative method based on image enhancement and feature fusion to reduce background noise and effectively handle the detection confusion caused by differences in ship sizes. Firstly, a feature-aware enhancement network is introduced, which preserves and strengthens the edge information of the target objects. Secondly, during the feature extraction phase, a dynamic hierarchical extraction module is proposed, significantly improving the feature capture ability of convolutional neural networks and overcoming the limitations of traditional fixed kernel receptive fields. Finally, a feature fusion module based on attention gating is employed to fully leverage the complementary information between the original and enhanced images, achieving precise modeling and efficient fusion of inter-feature correlations. The proposed method is integrated with the YOLOv8 detection framework for target detection. Experimental results in the publicly available SSDD and HRSID datasets demonstrate detection accuracies of 97.9% and 93.2%, respectively, thus validating the superiority and robustness of the proposed method.
- 20 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications Letters
- Vol. 15 (21622337) , 285-289
A TDOA Localization Method of HF Source Based on Dual-Branch CNN and Transfer Learning
High-frequency (HF) source localization with time-difference-of-arrival technology is a challenge in the field of radio monitoring. To reduce the impact of time-varying ionosphere and complicated noise, this letter proposes a dual-branch convolutional neural network, combined with transfer learning to handle real-world scenarios. The network is firstly pretrained with a large amount of simulated data to extract and fuse multi-dimensional feature. Then transfer learning is used to fine-tune a part of the model parameters with very little actual data for real-world conditions. Both simulated and experimental results show that the proposed method outperforms the existing HF localization method, with the locating accuracy of 14.37 km and 34.26 km for simulation and experiment, respectively.
- 20 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications Letters
- Vol. 15 (21622337) , 285-289
A TDOA Localization Method of HF Source Based on Dual-Branch CNN and Transfer Learning
High-frequency (HF) source localization with time-difference-of-arrival technology is a challenge in the field of radio monitoring. To reduce the impact of time-varying ionosphere and complicated noise, this letter proposes a dual-branch convolutional neural network, combined with transfer learning to handle real-world scenarios. The network is firstly pretrained with a large amount of simulated data to extract and fuse multi-dimensional feature. Then transfer learning is used to fine-tune a part of the model parameters with very little actual data for real-world conditions. Both simulated and experimental results show that the proposed method outperforms the existing HF localization method, with the locating accuracy of 14.37 km and 34.26 km for simulation and experiment, respectively.
- 13 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications Letters
- Vol. 15 (21622337) , 81-85
DDMNet: A Dual-Stream Network With Differential Temporal and Multiscale Spectral Features for Automatic Modulation Recognition
Automatic Modulation Recognition is a key technology in non-cooperative communications. With the growing complexity of electromagnetic environments, achieving high-precision and low-complexity modulation recognition remains a considerable challenge. Addressing the limitations inherent in existing methods for feature extraction and fusion, this letter introduces a dual-stream network model for modulation classification. This model innovatively incorporates a parallel feature extraction architecture that leverages differential amplitude-phase time-domain features alongside multi-scale spectral features. Furthermore, it achieves deep feature fusion through a dynamic gating mechanism. Experimental results demonstrate that the proposed network achieves an average recognition accuracy of 65.32% and 66.13% on the RML2016.10a and RML2016.10b datasets, respectively, surpassing the state-of-the-art method by 1.53% and 0.66%. This advancement significantly enhances the recognition performance, particularly under low signal-to-noise ratios and for high-order QAM signals.
- 13 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications Letters
- Vol. 15 (21622337) , 81-85
DDMNet: A Dual-Stream Network With Differential Temporal and Multiscale Spectral Features for Automatic Modulation Recognition
Automatic Modulation Recognition is a key technology in non-cooperative communications. With the growing complexity of electromagnetic environments, achieving high-precision and low-complexity modulation recognition remains a considerable challenge. Addressing the limitations inherent in existing methods for feature extraction and fusion, this letter introduces a dual-stream network model for modulation classification. This model innovatively incorporates a parallel feature extraction architecture that leverages differential amplitude-phase time-domain features alongside multi-scale spectral features. Furthermore, it achieves deep feature fusion through a dynamic gating mechanism. Experimental results demonstrate that the proposed network achieves an average recognition accuracy of 65.32% and 66.13% on the RML2016.10a and RML2016.10b datasets, respectively, surpassing the state-of-the-art method by 1.53% and 0.66%. This advancement significantly enhances the recognition performance, particularly under low signal-to-noise ratios and for high-order QAM signals.
- 1 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/CAA Journal of Automatica Sinica
- Vol. 12 (4) , 789-805
A Multi-Type Feature Fusion Network Based on Importance Weighting for Occluded Human Pose Estimation
Human pose estimation is a challenging task in computer vision. Most algorithms perform well in regular scenes, but lack good performance in occlusion scenarios. Therefore, we propose a multi-type feature fusion network based on importance weighting, which consists of three modules. In the first module, we propose a multi-resolution backbone with two feature enhancement sub-modules, which can extract features from different scales and enhance the feature expression ability. In the second module, we enhance the expressiveness of keypoint features by suppressing obstacle features and compensating for the unique and shared attributes of keypoints and topology. In the third module, we perform importance weighting on the adjacency matrix to enable it to describe the correlation among nodes, thereby improving the feature extraction ability. We conduct comparative experiments on the keypoint detection datasets of common objects in Context 2017 (COCO2017), COCO-Whole-body and CrowdPose, achieving the accuracy of 78.9%, 67.1% and 77.6%, respectively. Additionally, a series of ablation experiments are designed to show the performance of our work. Finally, we present the visualization of different scenarios to verify the effectiveness of our work.
- 1 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/CAA Journal of Automatica Sinica
- Vol. 12 (4) , 789-805
A Multi-Type Feature Fusion Network Based on Importance Weighting for Occluded Human Pose Estimation
Human pose estimation is a challenging task in computer vision. Most algorithms perform well in regular scenes, but lack good performance in occlusion scenarios. Therefore, we propose a multi-type feature fusion network based on importance weighting, which consists of three modules. In the first module, we propose a multi-resolution backbone with two feature enhancement sub-modules, which can extract features from different scales and enhance the feature expression ability. In the second module, we enhance the expressiveness of keypoint features by suppressing obstacle features and compensating for the unique and shared attributes of keypoints and topology. In the third module, we perform importance weighting on the adjacency matrix to enable it to describe the correlation among nodes, thereby improving the feature extraction ability. We conduct comparative experiments on the keypoint detection datasets of common objects in Context 2017 (COCO2017), COCO-Whole-body and CrowdPose, achieving the accuracy of 78.9%, 67.1% and 77.6%, respectively. Additionally, a series of ablation experiments are designed to show the performance of our work. Finally, we present the visualization of different scenarios to verify the effectiveness of our work.
- Open Access
- 11 November 2024
- Published by Springer Nature in Complex & Intelligent Systems
- Vol. 11 (1) , 1-15
PCNet: a human pose compensation network based on incremental learning for sports actions estimation
- 11 November 2024
- Published by Springer Nature in Complex & Intelligent Systems
- Vol. 11 (1) , 1-15
Open AccessPCNet: a human pose compensation network based on incremental learning for sports actions estimation
- 26 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 28 (2) , 303-307
TDOA Based Direct Positioning of Co-Channel Signals Using Cyclic Cross-Spectral Functions
Passive positioning of co-channel transmitters is a major challenge in radio monitoring. Co-channel signals overlap in both the time and frequency domains but have different cyclic frequencies. A direct localization method of multiple co-channel signals is proposed in this letter based on the relative time delay information of collected signals and cyclic cross-spectral functions. It can estimate the number and cyclic frequencies of co-channel signals and separate co-channel signals through spectral slicing for direct positioning. A joint spectral function is constructed with a set of randomly generated position points, and the position estimate for each co-channel signal can be determined by searching for the maximum value of the function. Additionally, an iterative sampling method is proposed to improve the positioning accuracy with only a few additional iterations. The simulation results demonstrate the superiority of the proposed algorithm over other joint positioning algorithms in the presence of co-channel interference.
- 26 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Letters
- Vol. 28 (2) , 303-307
TDOA Based Direct Positioning of Co-Channel Signals Using Cyclic Cross-Spectral Functions
Passive positioning of co-channel transmitters is a major challenge in radio monitoring. Co-channel signals overlap in both the time and frequency domains but have different cyclic frequencies. A direct localization method of multiple co-channel signals is proposed in this letter based on the relative time delay information of collected signals and cyclic cross-spectral functions. It can estimate the number and cyclic frequencies of co-channel signals and separate co-channel signals through spectral slicing for direct positioning. A joint spectral function is constructed with a set of randomly generated position points, and the position estimate for each co-channel signal can be determined by searching for the maximum value of the function. Additionally, an iterative sampling method is proposed to improve the positioning accuracy with only a few additional iterations. The simulation results demonstrate the superiority of the proposed algorithm over other joint positioning algorithms in the presence of co-channel interference.