基于多任务学习的间质性肺病分割算法  

Interstitial lung disease segmentation algorithm based on multi-task learning

作　　者：李威 陈玲[2] 徐修远[1] 朱敏[2] 郭际香[1] 周凯 牛颢[1] 张煜宸 易珊烨 章毅[1] 罗凤鸣[2] LI Wei;CHEN Ling;XU Xiuyuan;ZHU Min;GUO Jixiang;ZHOU Kai;NIU Hao;ZHANG Yuchen;YI Shanye;ZHANG Yi;LUO Fengming(College of Computer Science,Sichuan University,Chengdu Sichuan 610065,China;West China Hospital,Sichuan University,Chengdu Sichuan 610044,China;West China School of Medicine,Sichuan University,Chengdu Sichuan 610041,China)

机构地区：[1]四川大学计算机学院,成都610065 [2]四川大学华西医院,成都610044 [3]四川大学华西临床医学院,成都610041

出　　处：《计算机应用》2024年第4期1285-1293,共9页journal of Computer Applications

基　　金：国家自然科学基金资助项目(62106163);中国人工智能学会-华为MindSpore学术奖励基金资助项目(21H1235)。

摘　　要：间质性肺病(ILD)的分割标签标注成本极高,且现有数据集通常存在样本量较少的问题,导致训练的模型效果较差。针对该问题,提出一种基于多任务学习的ILD分割算法。首先,基于U-Net构建多任务分割模型;其次,使用生成的肺部分割标签作为辅助任务标签进行多任务学习;最后,使用一种自适应调整多任务损失函数权重的方法,平衡主任务和辅助任务的损失。在自构建的ILD数据集上的实验结果表明,多任务分割模型的Dice相似系数(DSC)达到了82.61%,与U-Net相比提升了2.26个百分点。验证了所提算法可以提升ILD的分割性能,协助临床医生进行ILD诊断。Interstitial Lung Disease(ILD)segmentation labels are highly costly,leading to small sample sizes in existing datasets and resulting in poor performance of trained models.To address this issue,a segmentation algorithm for ILD based on multi-task learning was proposed.Firstly,a multi-task segmentation model was constructed based on U-Net.Then,the generated lung segmentation labels were used as auxiliary task labels for multi-task learning.Finally,a method of dynamically weighting the multi-task loss functions was used to balance the losses of the primary task and the secondary task.Experimental results on a self-built ILD dataset show that the Dice Similarity Coefficient(DSC)of the multi-task segmentation model reaches 82.61%,which is 2.26 percentage points higher than that of U-Net.The experimental results demonstrate that the proposed algorithm can improve the segmentation performance of ILD and can assist clinical doctors in ILD diagnosis.

关 键 词：间质性肺病 语义分割 小样本量 多任务学习 自适应多任务损失函数 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]