Boheng Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Boheng Zhang has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Mingjian Sun
Harbin Institute of Technology
7 shared publications - Haorui Huang
Harbin Institute of Technology
3 shared publications - Yi Shen
Harbin Institute of Technology
3 shared publications - Baiya Li
Department of Otolaryngology Head and Neck Surgery, the First Affiliated Hospital of Xi'an Jiaotong University, Xi'an, 710061, China
2 shared publications - Cailing Lin
Harbin Institute of Technology
2 shared publications - Rui Fan
College of Electronics and Information Engineering, Tongji University, Shanghai, China
2 shared publications
Latest Publications
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (10) , 15900-15916
FSUS-Seg: A Fusion-Spectral-Driven Universal Medical Ultrasound Image Segmentation Network
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (10) , 15900-15916
FSUS-Seg: A Fusion-Spectral-Driven Universal Medical Ultrasound Image Segmentation Network
- Open Access
- 29 July 2026
- Published by Elsevier in SSRN Electronic Journal
WG-Net: Wavelet-Guided Network for Ultrasound Medical Image Segmentation
Ultrasound image segmentation plays a crucial role in quantitative lesion analysis, yet remains highly challenging due to inherent noise, low contrast, and blurred boundaries. Most existing deep learning–based segmentation methods struggle to produce accurate results under these conditions. To address these challenges, we propose a wavelet-guided network, named WG-Net, for ultrasound image segmentation. In the encoder, we design a Wavelet Fusion Enhancement Module (WFEM) to achieve a joint spatial–frequency representation of local and global features, thereby enhancing salient structures while suppressing noise. At the bottleneck, we introduce a Wavelet-Guided Mamba (WGM) module to enable long-range contextual interactions between spatial- and frequency-domain deep features, thereby facilitating multi-scale and cross-domain fusion for precise foreground–background discrimination. Furthermore, a Residual Wavelet (RW) module is incorporated into the skip connections to preserve essential high-frequency details while maintaining a large receptive field. This design strengthens the feature representations passed to the decoder, enabling high-quality reconstruction. Extensive experiments on three public ultrasound image segmentation datasets demonstrate the high segmentation accuracy and strong generalization ability of the proposed WG-Net, achieving Dice scores of 73.07% on BUSI, 80.27% on DDTI, and 87.05% on UDIAT.
- 29 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessWG-Net: Wavelet-Guided Network for Ultrasound Medical Image Segmentation
Ultrasound image segmentation plays a crucial role in quantitative lesion analysis, yet remains highly challenging due to inherent noise, low contrast, and blurred boundaries. Most existing deep learning–based segmentation methods struggle to produce accurate results under these conditions. To address these challenges, we propose a wavelet-guided network, named WG-Net, for ultrasound image segmentation. In the encoder, we design a Wavelet Fusion Enhancement Module (WFEM) to achieve a joint spatial–frequency representation of local and global features, thereby enhancing salient structures while suppressing noise. At the bottleneck, we introduce a Wavelet-Guided Mamba (WGM) module to enable long-range contextual interactions between spatial- and frequency-domain deep features, thereby facilitating multi-scale and cross-domain fusion for precise foreground–background discrimination. Furthermore, a Residual Wavelet (RW) module is incorporated into the skip connections to preserve essential high-frequency details while maintaining a large receptive field. This design strengthens the feature representations passed to the decoder, enabling high-quality reconstruction. Extensive experiments on three public ultrasound image segmentation datasets demonstrate the high segmentation accuracy and strong generalization ability of the proposed WG-Net, achieving Dice scores of 73.07% on BUSI, 80.27% on DDTI, and 87.05% on UDIAT.
- 15 April 2026
- Published by Elsevier in Neural Networks
- Vol. 202, 108994
FDPMambaFuse: A frequency-domain and parallel Mamba-based model for multimodal medical image fusion
- 15 April 2026
- Published by Elsevier in Neural Networks
- Vol. 202, 108994
FDPMambaFuse: A frequency-domain and parallel Mamba-based model for multimodal medical image fusion
- 25 March 2026
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 12 (2)
A Multimodal Attention Fusion-Based Model for Pathological Voice Detection
Pathological voice detection provides a non-invasive and cost-effective approach for the early screening of laryngeal disorders. However, most existing methods rely on single-modal acoustic features and mainly focus on binary classification, which limits their ability to achieve fine-grained discrimination among multiple pathological voice categories. To address these limitations, this paper proposes a Multimodal Fusion Network (MFNet) with attention-based fusion for pathological voice classification. The proposed framework jointly models time-domain and frequency-domain information within a unified architecture. Specifically, SincNet is employed to extract task-relevant narrowband acoustic features directly from raw speech waveforms, while the Tunable Q-factor Wavelet Transform (TQWT) and SKNet are combined to learn multi-scale spectral representations from Mel-spectrograms. In addition, an Attentional Feature Fusion (AFF) module is introduced to adaptively integrate features from the two branches, thereby enhancing the discriminative capability of pathological voice representations. Comparative experiments on three public pathological voice datasets show that the proposed method outperforms representative baseline models and demonstrates strong robustness and generalization across different datasets. Furthermore, ablation studies further validate the effectiveness of the proposed architectural design. These results indicate that MFNet provides an effective solution for pathological voice classification and has promising potential for computer-aided voice disorder screening.
- 25 March 2026
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 12 (2)
A Multimodal Attention Fusion-Based Model for Pathological Voice Detection
Pathological voice detection provides a non-invasive and cost-effective approach for the early screening of laryngeal disorders. However, most existing methods rely on single-modal acoustic features and mainly focus on binary classification, which limits their ability to achieve fine-grained discrimination among multiple pathological voice categories. To address these limitations, this paper proposes a Multimodal Fusion Network (MFNet) with attention-based fusion for pathological voice classification. The proposed framework jointly models time-domain and frequency-domain information within a unified architecture. Specifically, SincNet is employed to extract task-relevant narrowband acoustic features directly from raw speech waveforms, while the Tunable Q-factor Wavelet Transform (TQWT) and SKNet are combined to learn multi-scale spectral representations from Mel-spectrograms. In addition, an Attentional Feature Fusion (AFF) module is introduced to adaptively integrate features from the two branches, thereby enhancing the discriminative capability of pathological voice representations. Comparative experiments on three public pathological voice datasets show that the proposed method outperforms representative baseline models and demonstrates strong robustness and generalization across different datasets. Furthermore, ablation studies further validate the effectiveness of the proposed architectural design. These results indicate that MFNet provides an effective solution for pathological voice classification and has promising potential for computer-aided voice disorder screening.
- 24 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 29 (7) , 5069-5082
MCBTNet: Multi-Feature Fusion CNN and Bi- Level Routing Attention Transformer-Based Medical Image Segmentation Network
Accurate medical image segmentation is crucial for precise diagnosis and treatment in clinical pathology analysis and surgical navigation. While Convolutional Neural Network (CNN)-based approaches excel in capturing and analyzing local features, they often lose key global context. Transformers, utilizing self-attention mechanisms, address this issue but often overlook localized and multi-scale features while also requiring significant computational resources. To integrate the advantages of CNNs and Transformers to achieve efficient and precise medical image segmentation, we propose a segmentation framework based on multi-feature fusion CNN and Bi-level Routing Attention Transformer (MCBTNet). MCBTNet integrates CNNs and Transformers within a U-shaped encoderdecoder architecture. This configuration not only extracts multi-scale features via the U-shaped structure but also efficiently captures global contextual information through the dynamic sparsity of the Bi-Level Routing Attention Transformer. Our novel Frequency-Channel-Spatial multidimensional attention mechanism is implemented on skip connections, enhancing segmentation accuracy and speed by maximizing multi-scale feature utilization. Finally, MCBTNet obtains the segmentation result by fusing the predictions of different scales. Experimental results on five public datasets demonstrate that MCBTNet outperforms state-of-the-art methods in Dice and HD metrics, with lower computational and memory requirements. The code will be available on https://github.com/670768312/MCBTNet.
- 24 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 29 (7) , 5069-5082
MCBTNet: Multi-Feature Fusion CNN and Bi- Level Routing Attention Transformer-Based Medical Image Segmentation Network
Accurate medical image segmentation is crucial for precise diagnosis and treatment in clinical pathology analysis and surgical navigation. While Convolutional Neural Network (CNN)-based approaches excel in capturing and analyzing local features, they often lose key global context. Transformers, utilizing self-attention mechanisms, address this issue but often overlook localized and multi-scale features while also requiring significant computational resources. To integrate the advantages of CNNs and Transformers to achieve efficient and precise medical image segmentation, we propose a segmentation framework based on multi-feature fusion CNN and Bi-level Routing Attention Transformer (MCBTNet). MCBTNet integrates CNNs and Transformers within a U-shaped encoderdecoder architecture. This configuration not only extracts multi-scale features via the U-shaped structure but also efficiently captures global contextual information through the dynamic sparsity of the Bi-Level Routing Attention Transformer. Our novel Frequency-Channel-Spatial multidimensional attention mechanism is implemented on skip connections, enhancing segmentation accuracy and speed by maximizing multi-scale feature utilization. Finally, MCBTNet obtains the segmentation result by fusing the predictions of different scales. Experimental results on five public datasets demonstrate that MCBTNet outperforms state-of-the-art methods in Dice and HD metrics, with lower computational and memory requirements. The code will be available on https://github.com/670768312/MCBTNet.
- 4 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 72 (4) , 498-514
MM-UKAN++: A Novel Kolmogorov–Arnold Network-Based U-Shaped Network for Ultrasound Image Segmentation
Ultrasound (US) imaging is an important and commonly used medical imaging modality. Accurate and fast automatic segmentation of regions of interest (ROI) in US images is essential for enhancing the efficiency of clinical and robot-assisted diagnosis. However, US images suffer from low contrast, fuzzy boundaries, and significant scale variations in ROI. Existing convolutional neural network (CNN)-based and Transformer-based methods struggle with model efficiency and explainability. To address these challenges, we introduce MM-UKAN++, a novel U-shaped network based on Kolmogorov-Arnold Networks (KANs). MM-UKAN++ leverages multi-level KAN layers as the encoder and decoder within the U-network architecture, and incorporates an innovative multi-dimensional attention mechanism to refine skip connections by weighting features from frequency-channel and spatial perspectives. Additionally, the network effectively integrates multi-scale information, fusing outputs from different scale decoders to generate precise segmentation predictions. MM-UKAN++ achieves higher segmentation accuracy with lower computational cost and outperforms other mainstream methods on several open-source datasets for US image segmentation tasks, including achieving 69.42% IoU, 81.30% Dice, and 3.31mm HD in the BUSI dataset with 3.17G Flops and 9.90M Parameters. The excellent performance on our automatic carotid artery US scanning and diagnostic system further proves the speed and accuracy of MM-UKAN++. Besides, the good performance in other medical image segmentation tasks reveals the promising applications of MM-UKAN++. The code is available on GitHub.
- 4 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 72 (4) , 498-514
MM-UKAN++: A Novel Kolmogorov–Arnold Network-Based U-Shaped Network for Ultrasound Image Segmentation
Ultrasound (US) imaging is an important and commonly used medical imaging modality. Accurate and fast automatic segmentation of regions of interest (ROI) in US images is essential for enhancing the efficiency of clinical and robot-assisted diagnosis. However, US images suffer from low contrast, fuzzy boundaries, and significant scale variations in ROI. Existing convolutional neural network (CNN)-based and Transformer-based methods struggle with model efficiency and explainability. To address these challenges, we introduce MM-UKAN++, a novel U-shaped network based on Kolmogorov-Arnold Networks (KANs). MM-UKAN++ leverages multi-level KAN layers as the encoder and decoder within the U-network architecture, and incorporates an innovative multi-dimensional attention mechanism to refine skip connections by weighting features from frequency-channel and spatial perspectives. Additionally, the network effectively integrates multi-scale information, fusing outputs from different scale decoders to generate precise segmentation predictions. MM-UKAN++ achieves higher segmentation accuracy with lower computational cost and outperforms other mainstream methods on several open-source datasets for US image segmentation tasks, including achieving 69.42% IoU, 81.30% Dice, and 3.31mm HD in the BUSI dataset with 3.17G Flops and 9.90M Parameters. The excellent performance on our automatic carotid artery US scanning and diagnostic system further proves the speed and accuracy of MM-UKAN++. Besides, the good performance in other medical image segmentation tasks reveals the promising applications of MM-UKAN++. The code is available on GitHub.
- 31 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 70 (9) , 961-974
Visual Perception and Convolutional Neural Network-Based Robotic Autonomous Lung Ultrasound Scanning Localization System
Under the situation of severe COVID-19 epidemic, lung ultrasound (LUS) has been proved to be an effective and convenient method to diagnose and evaluate the extent of respiratory disease. However, the traditional clinical ultrasound (US) scanning requires doctors not only to be in close contact with patients, but also to have rich experience. In order to alleviate the shortage of medical resources and reduce the work stress and risk of infection for doctors, we propose a visual perception and convolutional neural network(CNN) based robotic autonomous LUS scanning localization system to realize scanned target recognition, probe pose solution and movement, and the acquisition of US images. The LUS scanned targets are identified through the target segmentation and localization algorithm based on the improved CNN, which is using the depth camera to collect the image information; Furthermore, the method based on multi-scale compensation normal vector is used to solve the attitude of the probe; Finally, a position control strategy based on force feedback is designed to optimize the position and attitude of the probe, which can not only obtain high-quality US images but also ensure the safety of patients and the system. The results of human LUS scanning experiment verify the accuracy and feasibility of the system. The positioning accuracy of the scanned targets is 15.63±0.18 mm, and the distance accuracy and rotation angle accuracy of the probe position calculation are 6.38 ± 0.25 mm and 8.60 ± 2.29 degrees, respectively. More importantly, the obtained high-quality US images can clearly capture the main pathological features of the lung. The system is expected to be applied in clinical practice.
- 31 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
- Vol. 70 (9) , 961-974
Visual Perception and Convolutional Neural Network-Based Robotic Autonomous Lung Ultrasound Scanning Localization System
Under the situation of severe COVID-19 epidemic, lung ultrasound (LUS) has been proved to be an effective and convenient method to diagnose and evaluate the extent of respiratory disease. However, the traditional clinical ultrasound (US) scanning requires doctors not only to be in close contact with patients, but also to have rich experience. In order to alleviate the shortage of medical resources and reduce the work stress and risk of infection for doctors, we propose a visual perception and convolutional neural network(CNN) based robotic autonomous LUS scanning localization system to realize scanned target recognition, probe pose solution and movement, and the acquisition of US images. The LUS scanned targets are identified through the target segmentation and localization algorithm based on the improved CNN, which is using the depth camera to collect the image information; Furthermore, the method based on multi-scale compensation normal vector is used to solve the attitude of the probe; Finally, a position control strategy based on force feedback is designed to optimize the position and attitude of the probe, which can not only obtain high-quality US images but also ensure the safety of patients and the system. The results of human LUS scanning experiment verify the accuracy and feasibility of the system. The positioning accuracy of the scanned targets is 15.63±0.18 mm, and the distance accuracy and rotation angle accuracy of the probe position calculation are 6.38 ± 0.25 mm and 8.60 ± 2.29 degrees, respectively. More importantly, the obtained high-quality US images can clearly capture the main pathological features of the lung. The system is expected to be applied in clinical practice.