Yuncong Feng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yuncong Feng has a total of 46 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
8 shared publications - Qingan Yao
College of Computer Science and Engineering, ChangChun University of Technology, Changchun, China
3 shared publications - Wanru Liu
College of Computer Science and Engineering, Changchun University of Technology, Changchun 130012, China
3 shared publications Xiaoli Zhang
3 shared publications- Cuixing Zhao
College of Computer Science and Engineering, Changchun University of Technology, Changchun 130012, China
2 shared publications - Guishen Wang
Changchun University of Technology
2 shared publications
Latest Publications
- 29 October 2026
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 82 (16) , 818
DAP-RTDETR: a small object detection algorithm based on Haar wavelet pyramid for UAV aerial images
- 29 October 2026
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 82 (16) , 818
DAP-RTDETR: a small object detection algorithm based on Haar wavelet pyramid for UAV aerial images
- Open Access
- 16 July 2026
- Published by Elsevier in SSRN Electronic Journal
AFM-HiFormer: Medical Image Segmentation Based on Adaptive Fusion and Multi-Enhanced Hybrid Transformer
Medical image segmentation technology has become an important part of auxiliary diagnosis and lesion detection. However, existing convolutional neural networks and Transformer models still suffer from problems such as over-reliance on local features, blurred boundaries, and insufficient generalization in small sample scenarios. In this paper, a novel medical image segmentation network based on adaptive fusion and multi-enhanced hybrid Transformer (AFM-HiFormer) is proposed. By constructing a cross-scale heterogeneous feature interaction and adaptive fusion mechanism, the model achieves dynamic fusion of local and global features. The method realizes dynamic weight learning of multi-level features by designing adaptive multi-scale fusion modules combined with dilated convolution and feature enhancement mechanisms; introduces boundary perception and deep supervision mechanisms to improve segmentation accuracy for small organs and complex structures; and employs multiple data enhancement modules that integrate geometric, intensity, and other transformation strategies to enhance the generalization ability of the model. Experimental results on the Synapse dataset and the ACDC dataset show that the AFM-HiFormer model improves the Dice coefficient by about 4\%--5\% compared with mainstream methods and reduces the HD95 index by approximately 7\%, which fully verifies its advantages in the collaborative optimization of structural representation and data enhancement. This work provides a new idea for efficient medical image segmentation.
- 16 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessAFM-HiFormer: Medical Image Segmentation Based on Adaptive Fusion and Multi-Enhanced Hybrid Transformer
Medical image segmentation technology has become an important part of auxiliary diagnosis and lesion detection. However, existing convolutional neural networks and Transformer models still suffer from problems such as over-reliance on local features, blurred boundaries, and insufficient generalization in small sample scenarios. In this paper, a novel medical image segmentation network based on adaptive fusion and multi-enhanced hybrid Transformer (AFM-HiFormer) is proposed. By constructing a cross-scale heterogeneous feature interaction and adaptive fusion mechanism, the model achieves dynamic fusion of local and global features. The method realizes dynamic weight learning of multi-level features by designing adaptive multi-scale fusion modules combined with dilated convolution and feature enhancement mechanisms; introduces boundary perception and deep supervision mechanisms to improve segmentation accuracy for small organs and complex structures; and employs multiple data enhancement modules that integrate geometric, intensity, and other transformation strategies to enhance the generalization ability of the model. Experimental results on the Synapse dataset and the ACDC dataset show that the AFM-HiFormer model improves the Dice coefficient by about 4\%--5\% compared with mainstream methods and reduces the HD95 index by approximately 7\%, which fully verifies its advantages in the collaborative optimization of structural representation and data enhancement. This work provides a new idea for efficient medical image segmentation.
- Open Access
- 22 April 2026
- Published by Elsevier in SSRN Electronic Journal
SFANet: Medical Image Segmentation Network Based on Spatial and Frequency Attention
Medical image segmentation plays a crucial role in medical image analysis but remains challenging due to complex anatomical structures and ambiguous tissue boundaries. While existing approaches have limitations in effectively exploiting feature maps through joint spatial and frequency-domain representations, we proposes a novel medical image segmentation framework based on spatial and frequency attention mechanisms, namely SFANet, to enhance segmentation accuracy and robustness. We innovatively proposes a convolutional multi-head self-attention module (CMSA), a multi-scale hybrid convolution module module (MSHCB), a frequency-enhanced channel attention module (FECA) and a dynamic grouped convolutional spatial attention module (DGCSA). Specifically, the CMSA module enhances global feature dependency modeling while maintaining computational efficiency through convolutional operations. The FECA module further improves detail sensitivity by fusing multi-frequency features across different scales. The DGCSA module combines dynamic grouped convolution with spatial attention mechanisms to enhance feature extraction, while enabling cross-scale feature fusion through adaptive convolution kernel weight adjustment. To enhance multi-scale feature extraction, we designed a multi-scale hybrid convolutional block (MSHCB) that adaptively processes features across different spatial dimensions. We conducte experiments on seven different modality datasets, and experimental results demonstrate that the proposed framework achieves significant performance improvements in multiple medical image segmentation tasks, exhibiting stronger segmentation capabilities and higher robustness than traditional methods.
- 22 April 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessSFANet: Medical Image Segmentation Network Based on Spatial and Frequency Attention
Medical image segmentation plays a crucial role in medical image analysis but remains challenging due to complex anatomical structures and ambiguous tissue boundaries. While existing approaches have limitations in effectively exploiting feature maps through joint spatial and frequency-domain representations, we proposes a novel medical image segmentation framework based on spatial and frequency attention mechanisms, namely SFANet, to enhance segmentation accuracy and robustness. We innovatively proposes a convolutional multi-head self-attention module (CMSA), a multi-scale hybrid convolution module module (MSHCB), a frequency-enhanced channel attention module (FECA) and a dynamic grouped convolutional spatial attention module (DGCSA). Specifically, the CMSA module enhances global feature dependency modeling while maintaining computational efficiency through convolutional operations. The FECA module further improves detail sensitivity by fusing multi-frequency features across different scales. The DGCSA module combines dynamic grouped convolution with spatial attention mechanisms to enhance feature extraction, while enabling cross-scale feature fusion through adaptive convolution kernel weight adjustment. To enhance multi-scale feature extraction, we designed a multi-scale hybrid convolutional block (MSHCB) that adaptively processes features across different spatial dimensions. We conducte experiments on seven different modality datasets, and experimental results demonstrate that the proposed framework achieves significant performance improvements in multiple medical image segmentation tasks, exhibiting stronger segmentation capabilities and higher robustness than traditional methods.
- 20 December 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 189
A hybrid transformer network with residual-like connections for medical image segmentation
- 20 December 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 189
A hybrid transformer network with residual-like connections for medical image segmentation
- 27 June 2024
- Published by Elsevier in Knowledge-Based Systems
- Vol. 300
GCFormer: Multi-scale feature plays a crucial role in medical images segmentation
- 27 June 2024
- Published by Elsevier in Knowledge-Based Systems
- Vol. 300
GCFormer: Multi-scale feature plays a crucial role in medical images segmentation
- 11 June 2024
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83 (37) , 85331-85355
An image segmentation fusion algorithm based on density peak clustering and Markov random field
- 11 June 2024
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83 (37) , 85331-85355
An image segmentation fusion algorithm based on density peak clustering and Markov random field
- 1 April 2024
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 94
PAMSNet: A medical image segmentation network based on spatial pyramid and attention mechanism
- 1 April 2024
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 94
PAMSNet: A medical image segmentation network based on spatial pyramid and attention mechanism
- Open Access
- 21 December 2023
- Published by Frontiers Media SA in Frontiers in Plant Science
- Vol. 14, 1320448
Attention-aided lightweight networks friendly to smart weeding robot hardware resources for crops and weeds semantic segmentation
Weed control is a global issue of great concern, and smart weeding robots equipped with advanced vision algorithms can perform efficient and precise weed control. Furthermore, the application of smart weeding robots has great potential for building environmentally friendly agriculture and saving human and material resources. However, most networks used in intelligent weeding robots tend to solely prioritize enhancing segmentation accuracy, disregarding the hardware constraints of embedded devices. Moreover, generalized lightweight networks are unsuitable for crop and weed segmentation tasks. Therefore, we propose an Attention-aided lightweight network for crop and weed semantic segmentation. The proposed network has a parameter count of 0.11M, Floating-point Operations count of 0.24G. Our network is based on an encoder and decoder structure, incorporating attention module to ensures both fast inference speed and accurate segmentation while utilizing fewer hardware resources. The dual attention block is employed to explore the potential relationships within the dataset, providing powerful regularization and enhancing the generalization ability of the attention mechanism, it also facilitates information integration between channels. To enhance the local and global semantic information acquisition and interaction, we utilize the refinement dilated conv block instead of 2D convolution within the deep network. This substitution effectively reduces the number and complexity of network parameters and improves the computation rate. To preserve spatial information, we introduce the spatial connectivity attention block. This block not only acquires more precise spatial information but also utilizes shared weight convolution to handle multi-stage feature maps, thereby further reducing network complexity. The segmentation performance of the proposed network is evaluated on three publicly available datasets: the BoniRob dataset, the Rice Seeding dataset, and the WeedMap dataset. Additionally, we measure the inference time and Frame Per Second on the NVIDIA Jetson Xavier NX embedded system, the results are 18.14 msec and 55.1 FPS. Experimental results demonstrate that our network maintains better inference speed on resource-constrained embedded systems and has competitive segmentation performance.
- 21 December 2023
- Published by Frontiers Media SA in Frontiers in Plant Science
- Vol. 14, 1320448
Open AccessAttention-aided lightweight networks friendly to smart weeding robot hardware resources for crops and weeds semantic segmentation
Weed control is a global issue of great concern, and smart weeding robots equipped with advanced vision algorithms can perform efficient and precise weed control. Furthermore, the application of smart weeding robots has great potential for building environmentally friendly agriculture and saving human and material resources. However, most networks used in intelligent weeding robots tend to solely prioritize enhancing segmentation accuracy, disregarding the hardware constraints of embedded devices. Moreover, generalized lightweight networks are unsuitable for crop and weed segmentation tasks. Therefore, we propose an Attention-aided lightweight network for crop and weed semantic segmentation. The proposed network has a parameter count of 0.11M, Floating-point Operations count of 0.24G. Our network is based on an encoder and decoder structure, incorporating attention module to ensures both fast inference speed and accurate segmentation while utilizing fewer hardware resources. The dual attention block is employed to explore the potential relationships within the dataset, providing powerful regularization and enhancing the generalization ability of the attention mechanism, it also facilitates information integration between channels. To enhance the local and global semantic information acquisition and interaction, we utilize the refinement dilated conv block instead of 2D convolution within the deep network. This substitution effectively reduces the number and complexity of network parameters and improves the computation rate. To preserve spatial information, we introduce the spatial connectivity attention block. This block not only acquires more precise spatial information but also utilizes shared weight convolution to handle multi-stage feature maps, thereby further reducing network complexity. The segmentation performance of the proposed network is evaluated on three publicly available datasets: the BoniRob dataset, the Rice Seeding dataset, and the WeedMap dataset. Additionally, we measure the inference time and Frame Per Second on the NVIDIA Jetson Xavier NX embedded system, the results are 18.14 msec and 55.1 FPS. Experimental results demonstrate that our network maintains better inference speed on resource-constrained embedded systems and has competitive segmentation performance.
- 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
- 12 October 2023
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 88
A three-layer decomposition method based on structural texture perception for fusion of CT and MRI images
- 12 October 2023
- Published by Elsevier in Biomedical Signal Processing and Control
- Vol. 88
A three-layer decomposition method based on structural texture perception for fusion of CT and MRI images