Ningbo Industrial Vision and Industrial Intelligence Lab, Zhejiang Wanli University, Ningbo 315100, China.
Faculty of Information Science and Engineering, Ocean University of China, Qingdao 266100, China.
Tomography. 2024 Oct 1;10(10):1577-1590. doi: 10.3390/tomography10100116.
Existing methods for MRI brain tumor segmentation often suffer from excessive model parameters and suboptimal performance in delineating tumor boundaries.
For this issue, a lightweight MRI brain tumor segmentation method, enhanced by hierarchical feature fusion (EHFF), is proposed. This method reduces model parameters while improving segmentation performance by integrating hierarchical features. Initially, a fine-grained feature adjustment network is crafted and guided by global contextual information, leading to the establishment of an adaptive feature learning (AFL) module. This module captures the global features of MRI brain tumor images through macro perception and micro focus, adjusting spatial granularity to enhance feature details and reduce computational complexity. Subsequently, a hierarchical feature weighting (HFW) module is constructed. This module extracts multi-scale refined features through multi-level weighting, enhancing the detailed features of spatial positions and alleviating the lack of attention to local position details in macro perception. Finally, a hierarchical feature retention (HFR) module is designed as a supplementary decoder. This module retains, up-samples, and fuses feature maps from each layer, thereby achieving better detail preservation and reconstruction.
Experimental results on the BraTS 2021 dataset demonstrate that the proposed method surpasses existing methods. Dice similarity coefficients (DSC) for the three semantic categories ET, TC, and WT are 88.57%, 91.53%, and 93.09%, respectively.
现有的 MRI 脑肿瘤分割方法往往存在模型参数过多和肿瘤边界分割性能不佳的问题。
针对这一问题，提出了一种基于分层特征融合（EHFF）的轻量级 MRI 脑肿瘤分割方法。该方法通过整合分层特征来减少模型参数，同时提高分割性能。首先，通过全局上下文信息指导，设计了一个细粒度特征调整网络，从而建立了自适应特征学习（AFL）模块。该模块通过宏观感知和微观聚焦捕捉 MRI 脑肿瘤图像的全局特征，调整空间粒度以增强特征细节并降低计算复杂度。随后，构建了一个分层特征加权（HFW）模块。该模块通过多级加权提取多尺度细化特征，增强空间位置的细节特征，并减轻宏观感知中对局部位置细节的关注不足。最后，设计了一个分层特征保留（HFR）模块作为补充解码器。该模块保留、上采样和融合来自每个层的特征图，从而实现更好的细节保留和重建。
在 BraTS 2021 数据集上的实验结果表明，该方法优于现有方法。三个语义类别 ET、TC 和 WT 的 Dice 相似系数（DSC）分别为 88.57%、91.53%和 93.09%。
