Qiyuan Zheng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Qiyuan Zheng has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
- Yan Zhang
School of Electrical Engineering and Automation, Harbin Institute of Technology, Harbin 150001, China
2 shared publications - Anran Zhong
School of Electrical Engineering and Automation, Harbin Institute of Technology, Harbin, Heilongjiang, China
1 shared publication - B. Hu1 shared publication
- Chunlai Zhou
Renmin University of China
1 shared publication - Chunling Yang
School of Electrical Engineering and Automation, Harbin Institute of Technology, Harbin, Heilongjiang, China
1 shared publication - Haojie Deng
School of Electrical Engineering and Automation, Harbin Institute of Technology, China
1 shared publication
Latest Publications
- 9 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-15
FDPF-Net: A Full-Scale Dynamic Pyramid Fusion Network for Infrared Small Target Detection
Infrared small target detection (IRSTD) methods have been extensively researched for various military and civilian applications and have greatly developed with the progress of deep learning in recent years. However, the performance of IRSTD remains limited due to challenges such as weak detection capabilities for diverse target boundaries and the complex background clutter present in infrared images across different scenarios. To overcome these challenges, this paper proposes a two-stage end-to-end full-scale dynamic pyramid fusion network (FDPF-Net). This network aims to refine small target boundary information and enhance both background consistency and the contrast between the target and its surroundings. The FDPF-Net introduces a feature extraction trunk sub-network and a full-scale dynamic refinement sub-network to extract and refine target and background information. Additionally, the proposed cross-layer scale adaptive module which is positioned between the trunk and the refinement sub-networks, adaptively integrates and optimizes the full-scale feature representation for boundary feature compensation. Finally, a feature pyramid fusion module is used to fuse and exploit the intrinsic information of small targets, avoiding feature dilution during the information passing process. Experimental results on three public datasets demonstrate that the proposed FDPF-Net outperforms other state-of-the-art methods in terms of Intersection over Union (IoU), Dice Similarity Coefficient (DSC), Precision (Pre), and Sensitivity (Se) and also exhibits more robust segmentation performance. It also maintains a balance between segmentation performance and model complexity, indicating its significant potential for real-world IRSTD applications.
- 9 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-15
FDPF-Net: A Full-Scale Dynamic Pyramid Fusion Network for Infrared Small Target Detection
Infrared small target detection (IRSTD) methods have been extensively researched for various military and civilian applications and have greatly developed with the progress of deep learning in recent years. However, the performance of IRSTD remains limited due to challenges such as weak detection capabilities for diverse target boundaries and the complex background clutter present in infrared images across different scenarios. To overcome these challenges, this paper proposes a two-stage end-to-end full-scale dynamic pyramid fusion network (FDPF-Net). This network aims to refine small target boundary information and enhance both background consistency and the contrast between the target and its surroundings. The FDPF-Net introduces a feature extraction trunk sub-network and a full-scale dynamic refinement sub-network to extract and refine target and background information. Additionally, the proposed cross-layer scale adaptive module which is positioned between the trunk and the refinement sub-networks, adaptively integrates and optimizes the full-scale feature representation for boundary feature compensation. Finally, a feature pyramid fusion module is used to fuse and exploit the intrinsic information of small targets, avoiding feature dilution during the information passing process. Experimental results on three public datasets demonstrate that the proposed FDPF-Net outperforms other state-of-the-art methods in terms of Intersection over Union (IoU), Dice Similarity Coefficient (DSC), Precision (Pre), and Sensitivity (Se) and also exhibits more robust segmentation performance. It also maintains a balance between segmentation performance and model complexity, indicating its significant potential for real-world IRSTD applications.
- Open Access
- 24 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 50-56
Motor Imagery EEG Classification Based on CEEMDAN-CWT Characterization
Brain-computer interface (BCI) is a technology that uses EEG signals to realize human-computer interaction. Motor imagery is a commonly used EEG paradigm, which has the advantage of active control and can be used in neurorehabilitation, prosthetic control and other fields. Based on the BCIC IV 2b binary classification dataset, this paper studies the EEG preprocessing and time-frequency representation algorithms. We designed the CEEMDAN-CWT method to reduce the complexity of the algorithm by reducing the number of channels, using sample entropy and the K-means clustering method. By comparing various time-frequency representation methods, the proposed method exhibits a good ability to improve feature separability. The experimental results showed that the optimized CEEMDAN-CWT method improves the MI-EEG classification accuracy and stability. It could reach a higher accuracy and a lower standard deviation on the 9 subjects compared with the CWT time-frequency representation method.
- 24 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 50-56
Open AccessMotor Imagery EEG Classification Based on CEEMDAN-CWT Characterization
Brain-computer interface (BCI) is a technology that uses EEG signals to realize human-computer interaction. Motor imagery is a commonly used EEG paradigm, which has the advantage of active control and can be used in neurorehabilitation, prosthetic control and other fields. Based on the BCIC IV 2b binary classification dataset, this paper studies the EEG preprocessing and time-frequency representation algorithms. We designed the CEEMDAN-CWT method to reduce the complexity of the algorithm by reducing the number of channels, using sample entropy and the K-means clustering method. By comparing various time-frequency representation methods, the proposed method exhibits a good ability to improve feature separability. The experimental results showed that the optimized CEEMDAN-CWT method improves the MI-EEG classification accuracy and stability. It could reach a higher accuracy and a lower standard deviation on the 9 subjects compared with the CWT time-frequency representation method.
- 20 October 2022
- Published by Elsevier in Leukemia Research
- Vol. 121, S66-S67
ALK INHIBITOR PLUS VINBLASTINE FOR REFRACTORY/RELAPSED PEDIATRIC ALK+ ANAPLASTIC LARGE CELL LYMPHOMA: A PROSPECTIVE, ONE-ARM, OPEN-LABEL REAL-WORLD STUDY
- 20 October 2022
- Published by Elsevier in Leukemia Research
- Vol. 121, S66-S67
ALK INHIBITOR PLUS VINBLASTINE FOR REFRACTORY/RELAPSED PEDIATRIC ALK+ ANAPLASTIC LARGE CELL LYMPHOMA: A PROSPECTIVE, ONE-ARM, OPEN-LABEL REAL-WORLD STUDY