Journal of Computer Applications ›› 2024, Vol. 44 ›› Issue (6): 1942-1948.DOI: 10.11772/j.issn.1001-9081.2023060742

Special Issue: 多媒体计算与计算机仿真

• Multimedia computing and computer simulation • Previous Articles     Next Articles

Rectified cross pseudo supervision method with attention mechanism for stroke lesion segmentation

Yan ZHOU1, Yang LI2()   

  1. 1.School of Computer Science and Engineering,Changchun University of Technology,Changchun Jilin 130012,China
    2.Academy for Advanced Interdisciplinary Studies,Northeast Normal University,Changchun Jilin 130024,China
  • Received:2023-06-12 Revised:2023-08-11 Accepted:2023-08-16 Online:2023-09-11 Published:2024-06-10
  • Contact: Yang LI
  • About author:ZHOU Yan, born in 1998, M. S. candidate. Her research interests include medical image analysis, deep learning.
  • Supported by:
    Jilin Provincial Science and Technology Development Foundation(20210201081GX)

用于脑卒中病灶分割的具有注意力机制的校正交叉伪监督方法

周妍1, 李阳2()   

  1. 1.长春工业大学 计算机科学与工程学院,长春 130012
    2.东北师范大学 前沿交叉研究院,长春 130024
  • 通讯作者: 李阳
  • 作者简介:周妍(1998—),女,江苏泰州人,硕士研究生,主要研究方向:医学影像分析、深度学习;
  • 基金资助:
    吉林省科技发展计划项目(20210201081GX)

Abstract:

The automatic segmentation of brain lesions provides a reliable basis for the timely diagnosis and treatment of stroke patients and the formulation of diagnosis and treatment plans, but obtaining large-scale labeled data is expensive and time-consuming. Semi-Supervised Learning (SSL) methods alleviate this problem by utilizing a large number of unlabeled images and a limited number of labeled images. Aiming at the two problems of pseudo-label noise in SSL and the lack of ability of existing Three-Dimensional (3D) networks to focus on smaller objects, a semi-supervised method was proposed, namely, a rectified cross pseudo supervised method with attention mechanism for stroke lesion segmentation RPE-CPS (Rectified Cross Pseudo Supervision with Project & Excite modules). First, the data was input into two 3D U-Net segmentation networks with the same structure but different initializations, and the obtained pseudo-segmentation graphs were used for cross-supervised training of the segmentation networks, making full use of the pseudo-label data to expand the training set, and encouraging a high similarity between the predictions of different initialized networks for the same input image. Second, a correction strategy about cross-pseudo-supervised approach based on uncertainty estimation was designed to reduce the impact of the noise in pseudo-labels. Finally, in the segmentation network of 3D U-Net, in order to improve the segmentation performance of small object classes, Project & Excite (PE) modules were added behind each encoder module, decoder module and bottleneck module. In order to verify the effectiveness of the proposed method, evaluation experiments were carried out on the Acute Ischemic Stroke (AIS) dataset of the cooperative hospital and the Ischemic Stroke Lesion Segmentation Challenge (ISLES2022) dataset. The experimental results showed that when only using 20% of the labeled data in the training set, the Dice Similarity Coefficient (DSC), 95% Hausdorff Distance (HD95), and Average Surface Distance (ASD) on the public dataset ISLES2022 reached 73.87%, 6.08 mm and 1.31 mm; on the AIS dataset, DSC, HD95, and ASD reached 67.74%, 15.38 mm and 1.05 mm, respectively. Compared with the state-of-the-art semi-supervised method Uncertainty Rectified Pyramid Consistency(URPC), DSC improved by 2.19 and 3.43 percentage points, respectively. The proposed method can effectively utilize unlabeled data to improve segmentation accuracy, outperforms other semi-supervised methods, and is robust.

Key words: medical image, stroke lesion segmentation, Semi-Supervised Learning (SSL), attention mechanism, uncertainty estimation

摘要:

脑部病变的自动分割为脑卒中患者的及时诊治和诊疗方案的制定提供了可靠的依据,但获取大规模标记数据昂贵且耗时。半监督学习(SSL)方法通过利用大量的未标记图像和有限的标记图像缓解这一问题。针对SSL中伪标签存在噪声,以及现有的三维(3D)网络缺乏聚焦较小目标的能力这2个问题,提出一种半监督方法,即用于脑卒中病灶分割的具有注意力机制的校正交叉伪监督方法RPE-CPS(Rectified Cross Pseudo Supervision with Project & Excite modules)。首先,将数据输入2个结构相同但初始化不同的3D U-Net分割网络,将得到的伪分割图用于交叉监督训练分割网络,充分利用伪标签数据扩展训练集,并鼓励不同初始化网络对同一输入图像的预测之间具有较高的相似性;其次,设计一种基于不确定性估计的交叉伪监督方法的校正策略,以降低伪标签中的噪声影响;最后,在3D U-Net分割网络中,为提高小目标类的分割性能,将投影-激发(PE)模块添加至每一个编码器模块、解码器模块和瓶颈模块之后。为验证所提方法的有效性,在合作医院急性缺血性脑卒中(AIS)数据集和缺血性脑卒中病灶分割挑战赛(ISLES2022)数据集上分别进行评估实验。实验结果表明,在仅使用训练集中20%的标记数据时,在公开数据集ISLES2022上Dice相似系数(DSC)、95%豪斯多夫距离(HD95)和平均表面距离(ASD)分别达到了73.87%、6.08 mm和1.31 mm;在AIS数据集上DSC、HD95和ASD分别达到了67.74%、15.38 mm和1.05 mm。与先进的半监督方法不确定性校正金字塔(URPC)相比,DSC分别提升了2.19和3.43个百分点。所提方法可以有效地利用未标记数据提高分割精度,优于其他半监督方法,并具有鲁棒性。

关键词: 医学影像, 脑卒中病灶分割, 半监督学习, 注意力机制, 不确定性估计

CLC Number: