Xiongfei Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xiongfei Li has a total of 79 co-authors. Here are the top co-authors by number of shared publications:
- Xiaoli Zhang
College of Computer Science and Technology, Jilin University, Changchun, China
15 shared publications - Kai Guo
Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun, China
4 shared publications - Siwei Ma
National Engineering Research Center of Visual Technology, Peking University, Beijing, China
4 shared publications - Rui Zhu
College of Communication Engineering, Jilin University, Changchun, China
3 shared publications - Tiehu Fan
College of Instrumentation and Electrical Engineering, Jilin University, Changchun, China
3 shared publications Xiaoli Zhang
Jilin University
3 shared publications
Latest Publications
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64  (01962892) , 1-15
DeSPAR: Depth-Guided Semantic-Prompted Adaptive Refinement for ORSI Salient Object Detection
Optical remote sensing image salient object detection (ORSI-SOD) currently faces two major challenges: (1) existing RGB-based methods primarily depend on color and texture cues, which makes it difficult to obtain robust representations of object spatial structure under extreme imaging conditions; (2) significant morphological variations across object categories lead to semantic feature confusion in existing methods. To address these issues, we propose Depth-Guided Semantic-Prompted Adaptive Refinement (DeSPAR), a progressive refinement framework with geometric–semantic decoupling. To avoid excessive coupling between geometric and semantic signals in an end-to-end architecture, which would cause semantic priors to interfere too early with the construction of generic geometric representations, DeSPAR adopts a two-stage design for feature learning. In Stage 1, Depth-Guided Geometric Learning (DGL) employs a novel lightweight Depth-Guided Refiner (DGR) to build a generic geometric foundation. DGR utilizes RGB features to guide pseudo-depth denoising and injects geometric cues from pseudo-depth to enhance spatial feature representations. In Stage 2, Depth-Guided Semantic-Adaptive Refinement (DSR) inherits the encoder weights from DGL and introduces category-specific constraints. Under the guidance of a Semantic Prompt Bank constructed from DGL, DSR adaptively optimizes the representations of different categories through a prompt-guided mechanism. Experimental results demonstrate that DeSPAR surpasses 22 state-of-the-art methods on three public ORSI-SOD benchmarks, achieving superior performance with only 26.4M parameters while attaining an inference speed of 161 FPS.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64  (01962892) , 1-15
DeSPAR: Depth-Guided Semantic-Prompted Adaptive Refinement for ORSI Salient Object Detection
Optical remote sensing image salient object detection (ORSI-SOD) currently faces two major challenges: (1) existing RGB-based methods primarily depend on color and texture cues, which makes it difficult to obtain robust representations of object spatial structure under extreme imaging conditions; (2) significant morphological variations across object categories lead to semantic feature confusion in existing methods. To address these issues, we propose Depth-Guided Semantic-Prompted Adaptive Refinement (DeSPAR), a progressive refinement framework with geometric–semantic decoupling. To avoid excessive coupling between geometric and semantic signals in an end-to-end architecture, which would cause semantic priors to interfere too early with the construction of generic geometric representations, DeSPAR adopts a two-stage design for feature learning. In Stage 1, Depth-Guided Geometric Learning (DGL) employs a novel lightweight Depth-Guided Refiner (DGR) to build a generic geometric foundation. DGR utilizes RGB features to guide pseudo-depth denoising and injects geometric cues from pseudo-depth to enhance spatial feature representations. In Stage 2, Depth-Guided Semantic-Adaptive Refinement (DSR) inherits the encoder weights from DGL and introduces category-specific constraints. Under the guidance of a Semantic Prompt Bank constructed from DGL, DSR adaptively optimizes the representations of different categories through a prompt-guided mechanism. Experimental results demonstrate that DeSPAR surpasses 22 state-of-the-art methods on three public ORSI-SOD benchmarks, achieving superior performance with only 26.4M parameters while attaining an inference speed of 161 FPS.
- 4 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 28  (15209210) , 1390-1403
A Semantic-Aware and Multi-Guided Network for Infrared-Visible Image Fusion
Multi-modality image fusion aims at fusing modality-specific (complementarity) and modality-shared (correlation) information from multiple source images. To tackle the overlooking of inter-feature relationships, high-frequency information loss, and the limited attention to downstream tasks, this paper focuses on efficiently extracting complementary in formation and aggregating multi-guided features. We propose a three-branch encoder-decoder architecture along with corresponding fusion layers as the fusion strategy. Firstly, shallow features from individual modalities are extracted by a depthwise convolution layer combined with the transformer block. In the three parallel branches of the encoder, Cross Attention and Invertible Block (CAI) extracts local features and preserves high frequency texture details. Base Feature Extraction Module (BFE) captures long-range dependencies and enhances modality-shared information. Graph Reasoning Module (GR) is introduced to reason high-level cross-modality relations and simultaneously ex tract low-level detail features as CAI's modality-specific complementary information. Experiments demonstrate the competitive results compared with state-of-the-art methods in visible/infrared image fusion and medical image fusion tasks. Moreover, the proposed algorithm surpasses the state-of-the-art methods in terms of subsequent tasks, averagely scoring 8.27% mAP@0.5 higher in object detection and 5.85% mIoU higher in semantic segmentation.
- 4 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 28  (15209210) , 1390-1403
A Semantic-Aware and Multi-Guided Network for Infrared-Visible Image Fusion
Multi-modality image fusion aims at fusing modality-specific (complementarity) and modality-shared (correlation) information from multiple source images. To tackle the overlooking of inter-feature relationships, high-frequency information loss, and the limited attention to downstream tasks, this paper focuses on efficiently extracting complementary in formation and aggregating multi-guided features. We propose a three-branch encoder-decoder architecture along with corresponding fusion layers as the fusion strategy. Firstly, shallow features from individual modalities are extracted by a depthwise convolution layer combined with the transformer block. In the three parallel branches of the encoder, Cross Attention and Invertible Block (CAI) extracts local features and preserves high frequency texture details. Base Feature Extraction Module (BFE) captures long-range dependencies and enhances modality-shared information. Graph Reasoning Module (GR) is introduced to reason high-level cross-modality relations and simultaneously ex tract low-level detail features as CAI's modality-specific complementary information. Experiments demonstrate the competitive results compared with state-of-the-art methods in visible/infrared image fusion and medical image fusion tasks. Moreover, the proposed algorithm surpasses the state-of-the-art methods in terms of subsequent tasks, averagely scoring 8.27% mAP@0.5 higher in object detection and 5.85% mIoU higher in semantic segmentation.
- 14 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63  (01962892) , 1-18
BPDUN: Bidirectional Progressive Deep Unfolding Network for Pansharpening
Pansharpening integrates low-resolution multispectral (LRMS) images with high-resolution panchromatic (PAN) images to produce high-resolution multispectral (HRMS) images for subsequent tasks. Existing methods have improved the image quality but still suffer from spectral distortion and blurry effects due to the inherent problems of scale inconsistency and frequency mismatch between source images. Although some deep learning (DL)-based methods adopt the progressive fusion strategy to mitigate their impact, they remain opaque and fail to incorporate domain-specific knowledge, limiting generalization ability. To address these problems, we propose a novel bidirectional progressive deep unfolding network (BPDUN) for pansharpening by decoupling the fusion process into a multistage PAN-guided multispectral (MS) image restoration task. In the forward direction, each stage performs iterative restoration of MS images at a specific spatial scale. Each iteration includes a prior denoising block (PDB) and an image reconstruction block (IRB) derived from variational optimization (VO) formulas with clear physical interpretations. The designed PDB implements band-aware prior fusion and dual-domain denoising, while a content-guided pixel attention module (CGPAM) is proposed for detail refinement. In the reverse direction, PAN images undergo two levels of decomposition via wavelet transform. The low-frequency component serves as a spatial prior to guide detail recovery for MS images at the same scale, while the high-frequency components are modulated between stages via a detail-enhanced upsampling module (DEUM). The proposed DEUM performs adaptive high-frequency fine-tuning to ensure spectral fidelity while increasing spatial detail during upsampling. Extensive experiments demonstrate that our method outperforms other state-of-the-art (SOTA) algorithms in spectral and spatial fidelity.
- 14 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63  (01962892) , 1-18
BPDUN: Bidirectional Progressive Deep Unfolding Network for Pansharpening
Pansharpening integrates low-resolution multispectral (LRMS) images with high-resolution panchromatic (PAN) images to produce high-resolution multispectral (HRMS) images for subsequent tasks. Existing methods have improved the image quality but still suffer from spectral distortion and blurry effects due to the inherent problems of scale inconsistency and frequency mismatch between source images. Although some deep learning (DL)-based methods adopt the progressive fusion strategy to mitigate their impact, they remain opaque and fail to incorporate domain-specific knowledge, limiting generalization ability. To address these problems, we propose a novel bidirectional progressive deep unfolding network (BPDUN) for pansharpening by decoupling the fusion process into a multistage PAN-guided multispectral (MS) image restoration task. In the forward direction, each stage performs iterative restoration of MS images at a specific spatial scale. Each iteration includes a prior denoising block (PDB) and an image reconstruction block (IRB) derived from variational optimization (VO) formulas with clear physical interpretations. The designed PDB implements band-aware prior fusion and dual-domain denoising, while a content-guided pixel attention module (CGPAM) is proposed for detail refinement. In the reverse direction, PAN images undergo two levels of decomposition via wavelet transform. The low-frequency component serves as a spatial prior to guide detail recovery for MS images at the same scale, while the high-frequency components are modulated between stages via a detail-enhanced upsampling module (DEUM). The proposed DEUM performs adaptive high-frequency fine-tuning to ensure spectral fidelity while increasing spatial detail during upsampling. Extensive experiments demonstrate that our method outperforms other state-of-the-art (SOTA) algorithms in spectral and spatial fidelity.
- Open Access
Cross Spatial Temporal Fusion Attention for Remote Sensing Object Detection via Image Feature Matching
Effectively describing features for cross-modal remote sensing image matching remains a challenging task due to the significant geometric and radiometric differences between multimodal images. Existing methods primarily extract features at the fully connected layer but often fail to capture cross-modal similarities effectively. We propose a Cross Spatial Temporal Fusion (CSTF) mechanism that enhances feature representation by integrating scale-invariant keypoints detected independently in both reference and query images. Our approach improves feature matching in two ways: First, by creating correspondence maps that leverage information from multiple image regions simultaneously, and second, by reformulating the similarity matching process as a classification task using SoftMax and Fully Convolutional Network (FCN) layers. This dual approach enables CSTF to maintain sensitivity to distinctive local features while incorporating broader contextual information, resulting in robust matching across diverse remote sensing modalities. To demonstrate the practical utility of improved feature matching, we evaluate CSTF on object detection tasks using the HRSC2016 and DOTA benchmark datasets. Our method achieves state-of-theart performance with an average mAP of 90.99% on HRSC2016 and 90.86% on DOTA, outperforming existing models. The CSTF model maintains computational efficiency with an inference speed of 12.5 FPS. These results validate that our approach to crossmodal feature matching directly enhances downstream remote sensing applications such as object detection.
Open AccessCross Spatial Temporal Fusion Attention for Remote Sensing Object Detection via Image Feature Matching
Effectively describing features for cross-modal remote sensing image matching remains a challenging task due to the significant geometric and radiometric differences between multimodal images. Existing methods primarily extract features at the fully connected layer but often fail to capture cross-modal similarities effectively. We propose a Cross Spatial Temporal Fusion (CSTF) mechanism that enhances feature representation by integrating scale-invariant keypoints detected independently in both reference and query images. Our approach improves feature matching in two ways: First, by creating correspondence maps that leverage information from multiple image regions simultaneously, and second, by reformulating the similarity matching process as a classification task using SoftMax and Fully Convolutional Network (FCN) layers. This dual approach enables CSTF to maintain sensitivity to distinctive local features while incorporating broader contextual information, resulting in robust matching across diverse remote sensing modalities. To demonstrate the practical utility of improved feature matching, we evaluate CSTF on object detection tasks using the HRSC2016 and DOTA benchmark datasets. Our method achieves state-of-theart performance with an average mAP of 90.99% on HRSC2016 and 90.86% on DOTA, outperforming existing models. The CSTF model maintains computational efficiency with an inference speed of 12.5 FPS. These results validate that our approach to crossmodal feature matching directly enhances downstream remote sensing applications such as object detection.
- 6 March 2025
- Published by Springer Nature in International Journal of Computer Vision
- Vol. 133  (7) , 4442-4464
UMSCS: A Novel Unpaired Multimodal Image Segmentation Method Via Cross-Modality Generative and Semi-supervised Learning
- 6 March 2025
- Published by Springer Nature in International Journal of Computer Vision
- Vol. 133  (7) , 4442-4464
UMSCS: A Novel Unpaired Multimodal Image Segmentation Method Via Cross-Modality Generative and Semi-supervised Learning
- 22 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 407-412
An Online Task Assignment Method in the Dynamic Arrival Scenario of Participants
Task assignment is one of the critical issues in mobile crowdsensing (MCS) paradigm. For the dynamic arrival scenario of participants, an online task assignment method named OTAPDA, is proposed in the paper. The proposed method adopts a three-layer architecture with trustful edge nodes. To maximize the task completion rate in condition of preserving bidding privacy of participants, a confusion function based on the exponential mechanism of differential privacy is firstly designed at edge level. And then winner participants are selected according to historical completion rate and fuzzy bidding of the participants under silding window. Experimental results on a real dataset demonstrate the effectiveness of the proposed method.
- 22 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 407-412
An Online Task Assignment Method in the Dynamic Arrival Scenario of Participants
Task assignment is one of the critical issues in mobile crowdsensing (MCS) paradigm. For the dynamic arrival scenario of participants, an online task assignment method named OTAPDA, is proposed in the paper. The proposed method adopts a three-layer architecture with trustful edge nodes. To maximize the task completion rate in condition of preserving bidding privacy of participants, a confusion function based on the exponential mechanism of differential privacy is firstly designed at edge level. And then winner participants are selected according to historical completion rate and fuzzy bidding of the participants under silding window. Experimental results on a real dataset demonstrate the effectiveness of the proposed method.
- 30 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28  (2) , 858-869
MRL-Seg: Overcoming Imbalance in Medical Image Segmentation With Multi-Step Reinforcement Learning
Medical image segmentation is a critical task for clinical diagnosis and research. However, dealing with highly imbalanced data remains a significant challenge in medical image segmentation, where the region of interest (ROI) may exhibit substantial variations across different slices. This presents a significant hurdle to medical image segmentation, as conventional segmentation methods may either overlook the minority class or overly emphasize the majority class, ultimately leading to a decrease in the overall generalization ability of the segmentation results. To overcome this, we propose a novel approach based on multi-step reinforcement learning, which integrates prior knowledge of medical images and pixel-wise segmentation difficulty into the reward function. Our method treats each pixel as an individual agent, utilizing diverse actions to evaluate its relevance for segmentation. To validate the effectiveness of our approach, we conduct experiments on four imbalanced medical datasets, and the results show that our approach surpasses other state-of-the-art methods in highly imbalanced scenarios. These findings hold substantial implications for clinical diagnosis and research.
- 30 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28  (2) , 858-869
MRL-Seg: Overcoming Imbalance in Medical Image Segmentation With Multi-Step Reinforcement Learning
Medical image segmentation is a critical task for clinical diagnosis and research. However, dealing with highly imbalanced data remains a significant challenge in medical image segmentation, where the region of interest (ROI) may exhibit substantial variations across different slices. This presents a significant hurdle to medical image segmentation, as conventional segmentation methods may either overlook the minority class or overly emphasize the majority class, ultimately leading to a decrease in the overall generalization ability of the segmentation results. To overcome this, we propose a novel approach based on multi-step reinforcement learning, which integrates prior knowledge of medical images and pixel-wise segmentation difficulty into the reward function. Our method treats each pixel as an individual agent, utilizing diverse actions to evaluate its relevance for segmentation. To validate the effectiveness of our approach, we conduct experiments on four imbalanced medical datasets, and the results show that our approach surpasses other state-of-the-art methods in highly imbalanced scenarios. These findings hold substantial implications for clinical diagnosis and research.
- 31 October 2023
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83  (16) , 47919-47941
Electronic explosives inspection: a fine-grained X-ray benchmark and few-shot prohibited phone detection model
- 31 October 2023
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83  (16) , 47919-47941
Electronic explosives inspection: a fine-grained X-ray benchmark and few-shot prohibited phone detection model
- 4 July 2023
- Published by Springer Nature in Applied Intelligence
- Vol. 53  (19) , 22969-22990
Detecting sparse building change with ambiguous label using Siamese full-scale connected network and instance augmentation
- 4 July 2023
- Published by Springer Nature in Applied Intelligence
- Vol. 53  (19) , 22969-22990
Detecting sparse building change with ambiguous label using Siamese full-scale connected network and instance augmentation