基于教师督导的磁共振图像联合重建与分割  

Combinatorial Reconstruction and Segmentation of Magnetic Resonance Image Using Teacher Forcing

作　　者：张宇 李浩然 李程 李飞[1] 王珊珊[1] Zhang Yu;Li Haoran;Li Cheng;Li Fei;Wang Shanshan(Paul C.Lauterbur Research Center for Biomedical Imaging,Shenzhen Institute of Advanced Technology,Chinese Academy of Sciences,Shenzhen 518055,Guangdong,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院深圳先进技术研究院劳特伯生物医学成像研究中心,广东深圳518055 [2]中国科学院大学,北京100049

出　　处：《激光与光电子学进展》2022年第14期256-262,共7页Laser & Optoelectronics Progress

基　　金：国家自然科学基金(61871371,81830056);科技创新2030——“新一代人工智能”项目(2020AAA0104100,2020AAA0104105);广东省重点领域研发计划(2018B010109009);广东省科技计划(2020B1212060051);深圳市基础研究项目(JCYJ20180507182400762);中国科学院青年创新促进会项目(2019351)。

摘　　要：现有的深度学习方法倾向于将磁共振图像重建与分割作为两个单独的任务来处理,而没有考虑到两个任务之间的相关性。如果简单地对重建网络与分割网络进行拼接训练,则可能会由于任务之间的优化难度差异而影响重建与分割的最终效果。基于改进后的教师督导网络训练策略,开发了一种磁共振图像联合重建与分割的多任务深度学习方法。新设计的教师督导策略迭代地以中间重建输出和全采样数据来指导多任务网络训练,缓解误差积累。在一个公开数据集和一个内部数据集上对所提方法进行评估,并与6种现有方法进行了比较。实验结果表明,与对比方法相比,所提方法在实现重建与分割协同优化的同时提升了重建图像质量和分割精度。Existing deep learning methods handle magnetic resonance(MR)image reconstruction and segmentation as individual task instead of considering their correlations.However,the simple concatenation of the reconstruction and segmentation networks can compromise the performances on both tasks due to the differences in optimization.This paper develops a multi-task deep learning method for the combinatorial reconstruction and segmentation of MR images using an improved teacher forcing network training strategy.The newly designed teacher forcing scheme guides multi-task network training by iteratively using intermediate reconstruction outputs and fully sampled data to avoid error accumulation.We compared the effectiveness of the proposed method with six state-of-the-art methods on an open dataset and an in vivo inhouse dataset.The experimental results show that compared to other methods,the proposed method possesses encouraging capabilities to achieve better image reconstruction quality and segmentation accuracy while co-optimizing MR image reconstruction and segmentation simultaneously.

关 键 词：计算机视觉 磁共振图像重建与分割 多任务网络 任务驱动成像 教师督导 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]