Skip to main content
QUICK REVIEW

[论文解读] Dual Encoder Fusion U-Net (DEFU-Net) for Cross-manufacturer Chest X-ray Segmentation

Lipei Zhang, Aozhi Liu|arXiv (Cornell University)|Sep 11, 2020
COVID-19 diagnosis using AI参考文献 41被引用 4
一句话总结

该论文提出DEFU-Net,一种双路编码器融合U-Net架构,结合Inception模块与空洞卷积,并采用密集连接循环卷积块,以增强跨制造商胸部X光片分割中的空间与上下文特征提取。该模型在Montgomery和深圳医院的混合数据集上实现了最先进性能,在Dice、IOU、F1-score和AUC等多个指标上均优于U-Net、Res-U-Net、BCDU-Net、R2U-Net及注意力增强变体。

ABSTRACT

A number of methods based on deep learning have been applied to medical image segmentation and have achieved state-of-the-art performance. Due to the importance of chest x-ray data in studying COVID-19, there is a demand for state-of-the-art models capable of precisely segmenting soft tissue on the chest x-rays. The dataset for exploring best segmentation model is from Montgomery and Shenzhen hospital which had opened in 2014. The most famous technique is U-Net which has been used to many medical datasets including the Chest X-rays. However, most variant U-Nets mainly focus on extraction of contextual information and skip connections. There is still a large space for improving extraction of spatial features. In this paper, we propose a dual encoder fusion U-Net framework for Chest X-rays based on Inception Convolutional Neural Network with dilation, Densely Connected Recurrent Convolutional Neural Network, which is named DEFU-Net. The densely connected recurrent path extends the network deeper for facilitating contextual feature extraction. In order to increase the width of network and enrich representation of features, the inception blocks with dilation are adopted. The inception blocks can capture globally and locally spatial information from various receptive fields. At the same time, the two paths are fused by summing features, thus preserving the contextual and spatial information for decoding part. This multi-learning-scale model is benefiting in Chest X-ray dataset from two different manufacturers (Montgomery and Shenzhen hospital). The DEFU-Net achieves the better performance than basic U-Net, residual U-Net, BCDU-Net, R2U-Net and attention R2U-Net. This model has proved the feasibility for mixed dataset and approaches state-of-the-art. The source code for this proposed framework is public https://github.com/uceclz0/DEFU-Net.

研究动机与目标

  • 解决在不同影像制造商设备中分割胸部X光片软组织的挑战,其中领域偏移和图像质量差异影响模型泛化能力。
  • 通过同时增强空间感知与上下文理解,改进医学图像分割中的特征提取。
  • 开发一种鲁棒模型,能够处理来自不同来源(Montgomery和深圳医院)的混合数据集,且性能不下降。
  • 在分割精度上超越现有U-Net变体,尤其在易误分类的边缘区域与小区域像素上表现更优。
  • 通过编码器路径的架构创新,实现异质数据上的更快收敛与更好拟合。

提出的方法

  • DEFU-Net采用双路编码器:一路使用带有空洞卷积的Inception模块,以捕捉不同感受野下的多尺度空间特征。
  • 第二路采用密集连接循环卷积块(DCRC),以加深网络并增强层间特征传播。
  • 在解码前,通过逐元素相加方式融合两条编码器路径的特征,保留丰富的上下文与空间信息,实现精准分割。
  • 网络采用类似U-Net的跳跃连接,连接编码器与解码器路径,以保持空间分辨率与梯度流动。
  • 瓶颈层的特征图数量设置为与第四层编码器匹配,使计算成本减半。
  • 模型使用交叉熵损失结合早停策略与Adam优化,实现快速收敛与高性能。

实验结果

研究问题

  • RQ1与标准U-Net变体相比,双路编码器架构是否能提升在跨制造商胸部X光片数据集上的分割性能?
  • RQ2空洞Inception模块的引入是否能增强在低对比度、边缘模糊的医学图像中多尺度空间特征的提取?
  • RQ3密集连接循环块在编码器中在多大程度上提升了特征表示的深度与上下文建模能力?
  • RQ4通过双路特征相加实现的特征融合,对解码精度有何影响,尤其是在器官边界附近?
  • RQ5该模型是否能在无需微调的情况下,有效泛化至不同影像制造商的数据集?

主要发现

  • DEFU-Net在测试集上达到最高的Dice分数0.9667与F1-score 0.9619,优于U-Net、Res-U-Net、BCDU-Net、Incep-Res-U-Net、R2U-Net与Att-R2U-Net。
  • 在跨制造商泛化测试中,当在Shenzhen数据集上训练、Montgomery数据集上测试时,Dice分数达0.9227;当在Montgomery数据集上训练、Shenzhen数据集上测试时,Dice分数达0.9154,表现出强大鲁棒性。
  • 该模型收敛速度快,仅两轮训练后准确率即达0.9776,且在145轮训练中未出现过拟合。
  • 使用空洞Inception模块显著优于标准3×3卷积,尤其在捕捉全局与局部空间模式方面表现更优。
  • 可视化结果表明,DEFU-Net在边缘区域与小肺区更准确地对齐预测与真实标签,显著减少假阳性和假阴性。
  • 空洞Inception与DCRC模块的结合,使模型在应对领域偏移与跨制造商图像差异方面表现更优。

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。