Bowen Zhou

Bowen Zhou

Auto-generated profile, based on publicly available information

Bowen Zhou

Anhui Provincial International Joint Research Center for Advanced Technology in Medical Imaging, Anhui University, Hefei, China

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
16 December 2024 in IEEE Transactions on Radiation and Plasma Medical Sciences
Verified
Journal Article
Closed access
BZBowen Zhou
Bowen Zhou
XDXingbo Dong
Xingbo Dong
XZXiaowei Zhao
Xiaowei Zhao
CLChenglong Li
Chenglong Li
ZJZhe Jin
Zhe Jin
HWHuabin Wang
Huabin Wang
The hybrid architecture of CNNs and Transformers has gained popularity in medical image segmentation. However, in this hybrid architecture, the semantic gaps between multi-scale CNN and Transformer branch features hinder the segmentation performance. To address this issue, we propose a new pipeline for medical image segmentation, named FCT-Net. Firstly, we employ large kernel convolutions in CNN branch and shift-window self-attention mechanism in Transformer branch to construct a parallel three-branch architecture for efficiently capturing global and local information. Secondly, to better fuse CNN and Transformer features, we introduce the Bridge Fusion Module (BFM), to effectively extract local features and global representations at different semantic scales by integrating semantic information from different branches, thereby reducing the semantic gap between features. Lastly, to capture multi-scale information during the encoding process, we design the Multi-Scale Feature Compilation Module (MFCM) to adaptively fuse features from different stages of the encoder. Additionally, we introduce Residual Attention (RA) to enhance the features obtained after encoding, further boosting the network’s representational capacity. FCT-Net is evaluated on four different medical image segmentation benchmarks, achieving Dice scores of 83.56%, 90.87%, 92.21%, and 91.92% on COVID-19 Lung dataset, ISIC-2018, SegPC-2021, and ACDC, respectively, outperforming other state-of-the-art methods. Source code will be available at https://github.com/ZBW830/FCTNethttps://github.com/ZBW830/FCTNet.

21 May 2025 in IEEE Journal of Biomedical and Health Informatics
Verified
Journal Article
Closed access
XDXingbo Dong
Xingbo Dong
BZBowen Zhou
Bowen Zhou
CYChen Yin
Chen Yin
ILIman Yi Liao
Iman Yi Liao
ZJZhe Jin
Zhe Jin
BPBin Pu
Bin Pu
Medical image segmentation is critical for disease diagnosis, treatment planning, and prognosis assessment, yet the complexity and diversity of medical images pose significant challenges to accurate segmentation. While Convolutional Neural Networks capture local features and Vision Transformers excel in the global context, both struggle with efficient long-range dependency modeling. Inspired by Mamba's State Space Modeling efficiency, we propose ÆMMamba, a novel multi-scale feature extraction framework built on the Mamba backbone network. AÆMMamba integrates several innovative modules: the Efficient Fusion Bridge (EFB) module, which employs a bidirectional state-space model and attention mechanisms to fuse multi-scale features; the Edge-Aware Module (EAM), which enhances low-level edge representation using Sobel-based edge extraction; and the Boundary Sensitive Decoder (BSD), which leverages inverse attention and residual convolutional layers to handle cross-level complex boundaries. ÆMMamba achieves state-of-the-art performance across 8 medical segmentation datasets. On polyp segmentation datasets (Kvasir, ClinicDB, ColonDB, EndoScene, ETIS), it records the highest mDice and mIoU scores, outperforming methods like MADGNet and Swin-UMamba, with a standout mDice of 72.22 on ETIS, the most challenging dataset in this domain. For lung and breast segmentation, ÆMMamba surpasses competitors such as H2Former and SwinUnet, achieving Dice scores of 84.24 on BUSI and 79.83 on COVID-19 Lung. And on the LGG brain MRI dataset, ÆMMamba attains an mDice of 87.25 and an mIoU of 79.31, outperforming all compared methods. The source code will be released at https://github.com/xingbod/eMMamba.