Suting Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Suting Chen has a total of 104 co-authors. Here are the top co-authors by number of shared publications:
Chuang Zhang
Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology (CICAEET), Nanjing University of Information Science and Technology, Nanjing, China
4 shared publications- Chuang Zhang
Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology (CICAEET), Nanjing University of Information Science & Technology, Nanjing 210044, China
3 shared publications Dongwei Shao
Nanjing University of Information Science and Technology, School of Artificial Intelligence/School of Future Technology, Nanjing, China
3 shared publications- Huajun Liu
School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, China
3 shared publications - Mithun Mukherjee
Nanjing University of Information Science and Technology, Nanjing, China
3 shared publications - Xinyu Lu
School of Electronic and Information Engineering, Nanjing University of Information Science and Technology, Nanjing, China
3 shared publications
Latest Publications
- 6 October 2026
- Published by Elsevier in Journal of the Franklin Institute
- Vol. 363  (17)
OccuPoint2Roof: A progressive joint completion and reconstruction framework for 3D building roofs from occluded LiDAR point clouds
- 6 October 2026
- Published by Elsevier in Journal of the Franklin Institute
- Vol. 363  (17)
OccuPoint2Roof: A progressive joint completion and reconstruction framework for 3D building roofs from occluded LiDAR point clouds
- 10 August 2026
- Published by Elsevier in Neurocomputing
- Vol. 706
TF-Net: Video prediction method based on Fourier transform and edge semantic segmentation
- 10 August 2026
- Published by Elsevier in Neurocomputing
- Vol. 706
TF-Net: Video prediction method based on Fourier transform and edge semantic segmentation
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25, 23470-23486
Hierarchical Group Method for Over-the-Air Channel Calibration in Cell-Free Massive MIMO
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25, 23470-23486
Hierarchical Group Method for Over-the-Air Channel Calibration in Cell-Free Massive MIMO
- 22 October 2025
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 163
GMFIMamba: Remote sensing change detection based on group Mamba feature interaction
- 22 October 2025
- Published by Elsevier in Engineering Applications of Artificial Intelligence
- Vol. 163
GMFIMamba: Remote sensing change detection based on group Mamba feature interaction
- 1 October 2025
- Published by Springer Nature
- p. 67-78
TDRSolver: Confidentiality-Preserving Repair of Inconsistent Data in Temporal Knowledge Graphs Using Ant Colony Optimization
- 1 October 2025
- Published by Springer Nature
- p. 67-78
TDRSolver: Confidentiality-Preserving Repair of Inconsistent Data in Temporal Knowledge Graphs Using Ant Colony Optimization
- 1 September 2025
- Published by AIP Publishing in Journal of Renewable and Sustainable Energy
- Vol. 17  (5)
MLATD: Photovoltaic power prediction method based on multi-level attention and temporal decoupling
With the increasing share of photovoltaic (PV) energy in power systems, PV power output affected by factors, such as geographical distribution and meteorological conditions, exhibits complex spatiotemporal dynamics. Traditional forecasting methods face challenges in effectively modeling spatial correlations, capturing multiscale temporal features, and maintaining training efficiency in deep networks. To address these issues, this paper proposes a multi-level attention and time decoupled (MLATD) method for PV power forecasting, which deeply explores the underlying variation patterns in PV generation data from both spatial and temporal dimensions. First, a graph convolutional network is employed to model the spatial correlations in PV generation data. Then, a multi-level attention unit incorporating contextual attention mechanisms is designed as the basic module of a gated recurrent unit (GRU) network to capture the global temporal dependencies of PV data. Furthermore, a time decoupling unit is inserted between GRU layers to decompose the PV generation sequences into transient fluctuations and trend components, enabling multi-scale feature perception. Finally, a high speed channel is introduced to optimize the gradient backpropagation path. Experiments are conducted on real world PV generation datasets for power forecasting tasks, with model performance evaluated using mean absolute error, mean squared error, root mean squared error, and R-square (R2) metrics. The results demonstrate that MLATD significantly outperforms traditional models, with improvements in forecasting metrics ranging from 7.362% to 17.953%.
- 1 September 2025
- Published by AIP Publishing in Journal of Renewable and Sustainable Energy
- Vol. 17  (5)
MLATD: Photovoltaic power prediction method based on multi-level attention and temporal decoupling
With the increasing share of photovoltaic (PV) energy in power systems, PV power output affected by factors, such as geographical distribution and meteorological conditions, exhibits complex spatiotemporal dynamics. Traditional forecasting methods face challenges in effectively modeling spatial correlations, capturing multiscale temporal features, and maintaining training efficiency in deep networks. To address these issues, this paper proposes a multi-level attention and time decoupled (MLATD) method for PV power forecasting, which deeply explores the underlying variation patterns in PV generation data from both spatial and temporal dimensions. First, a graph convolutional network is employed to model the spatial correlations in PV generation data. Then, a multi-level attention unit incorporating contextual attention mechanisms is designed as the basic module of a gated recurrent unit (GRU) network to capture the global temporal dependencies of PV data. Furthermore, a time decoupling unit is inserted between GRU layers to decompose the PV generation sequences into transient fluctuations and trend components, enabling multi-scale feature perception. Finally, a high speed channel is introduced to optimize the gradient backpropagation path. Experiments are conducted on real world PV generation datasets for power forecasting tasks, with model performance evaluated using mean absolute error, mean squared error, root mean squared error, and R-square (R2) metrics. The results demonstrate that MLATD significantly outperforms traditional models, with improvements in forecasting metrics ranging from 7.362% to 17.953%.
- 1 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74  (00189456) , 1-15
CrackMDM: Masked Modeling on DCT Domain for Efficient Pavement Crack Segmentation
Existing pavement crack inspection methods heavily rely on a large amount of annotated samples and require overburdened computational power, which is not affordable on edge devices. Recent methods mostly focus on spatial features, which cannot effectively capture the long, continuous, tender, and thin road features. The masked image modeling (MIM) [1] is an effective way to rebuild the crack primitive features by masking strategy on unlabeled data to reduce the dependence on annotated data, and the convolution on frequency domain provides a flexible and efficient path to capture continuous and tender road features. Inspired by the thoughts of masked frequency modeling, we proposed a masked DCT-domain modeling strategy, named CrackMDM, for efficient pavement crack segmentation. Specifically, we propose a DCT-domain masked modeling method in the CrackMDM model which combines the advantages of separable convolutions and spectral convolutions in the DCT-domain to extract continuous and tender crack structures. Additionally, we introduce the self-supervised pre-training with a masking strategy in DCT-domain using unlabeled crack samples to build crack primitives and to fine-tune the encoder and decoder parameters on labeled crack data to refine crack features in the fine-tuning phase. The CrackMDM model is evaluated on three public benchmarks: CFD, YCD, and GAPs, and achieves state-of-the-art (SOTA) performance with superior inference speed. Codes are available at https://github.com/Jyuan357/CrackMDM.
- 1 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74  (00189456) , 1-15
CrackMDM: Masked Modeling on DCT Domain for Efficient Pavement Crack Segmentation
Existing pavement crack inspection methods heavily rely on a large amount of annotated samples and require overburdened computational power, which is not affordable on edge devices. Recent methods mostly focus on spatial features, which cannot effectively capture the long, continuous, tender, and thin road features. The masked image modeling (MIM) [1] is an effective way to rebuild the crack primitive features by masking strategy on unlabeled data to reduce the dependence on annotated data, and the convolution on frequency domain provides a flexible and efficient path to capture continuous and tender road features. Inspired by the thoughts of masked frequency modeling, we proposed a masked DCT-domain modeling strategy, named CrackMDM, for efficient pavement crack segmentation. Specifically, we propose a DCT-domain masked modeling method in the CrackMDM model which combines the advantages of separable convolutions and spectral convolutions in the DCT-domain to extract continuous and tender crack structures. Additionally, we introduce the self-supervised pre-training with a masking strategy in DCT-domain using unlabeled crack samples to build crack primitives and to fine-tune the encoder and decoder parameters on labeled crack data to refine crack features in the fine-tuning phase. The CrackMDM model is evaluated on three public benchmarks: CFD, YCD, and GAPs, and achieves state-of-the-art (SOTA) performance with superior inference speed. Codes are available at https://github.com/Jyuan357/CrackMDM.
- 21 June 2025
- Published by Springer Nature in Applied Intelligence
- Vol. 55  (11) , 1-20
MSTD: A multimodal vehicle trajectory prediction method based on multi-level spatiotemporal decoupling
- 21 June 2025
- Published by Springer Nature in Applied Intelligence
- Vol. 55  (11) , 1-20
MSTD: A multimodal vehicle trajectory prediction method based on multi-level spatiotemporal decoupling
- 12 May 2025
- Published by Springer Nature in Multimedia Systems
- Vol. 31  (4) , 1-26
SEGMTM: a spectrum prediction method based on enhanced graph convolution and multi-scale time decomposition
- 12 May 2025
- Published by Springer Nature in Multimedia Systems
- Vol. 31  (4) , 1-26
SEGMTM: a spectrum prediction method based on enhanced graph convolution and multi-scale time decomposition
- 6 February 2025
- Published by Elsevier in Digital Signal Processing
- Vol. 160
HDA-DGCN: Hierarchical data-driven aggregation network assisted dynamic graph convolutional framework for meteorological prediction
- 6 February 2025
- Published by Elsevier in Digital Signal Processing
- Vol. 160
HDA-DGCN: Hierarchical data-driven aggregation network assisted dynamic graph convolutional framework for meteorological prediction