Institute of Medical Robotics, School of Biomedical Engineering, Shanghai Jiao Tong University, No. 800 Dongchuan Road, 200240 Shanghai, China.
sitem Center for Translational Medicine and Biomedical Entrepreneurship, University of Bern, Bern, Switzerland.
Med Image Anal. 2022 Nov;82:102607. doi: 10.1016/j.media.2022.102607. Epub 2022 Aug 30.
Despite remarkable success of deep learning, distribution divergence remains a challenge that hinders the performance of many tasks in medical image analysis. Large distribution gap may deteriorate the knowledge transfer across different domains or feature subspaces. To achieve better distribution alignment, we propose a novel module named Instance to Prototype Earth Mover's Distance (I2PEMD), where shared class-specific prototypes are progressively learned to narrow the distribution gap across different domains or feature subspaces, and Earth Mover's Distance (EMD) is calculated to take into consideration the cross-class relationships during embedding alignment. We validate the effectiveness of the proposed I2PEMD on two different tasks: multi-modal medical image segmentation and semi-supervised classification. Specifically, in multi-modal medical image segmentation, I2PEMD is explicitly utilized as a distribution alignment regularization term to supervise the model training process, while in semi-supervised classification, I2PEMD works as an alignment measure to sort and cherry-pick the unlabeled data for more accurate and robust pseudo-labeling. Results from comprehensive experiments demonstrate the efficacy of the present method.
尽管深度学习取得了显著的成功，但分布差异仍然是一个挑战，它阻碍了医学图像分析中许多任务的性能。较大的分布差距可能会恶化不同领域或特征子空间之间的知识转移。为了实现更好的分布对齐，我们提出了一个名为实例到原型 Earth Mover's Distance（I2PEMD）的新模块，其中逐步学习共享的特定于类的原型，以缩小不同领域或特征子空间之间的分布差距，并计算 Earth Mover's Distance（EMD），以在嵌入对齐过程中考虑跨类关系。我们在两个不同的任务上验证了所提出的 I2PEMD 的有效性：多模态医学图像分割和半监督分类。具体来说，在多模态医学图像分割中，I2PEMD 被明确用作分布对齐正则化项来监督模型训练过程，而在半监督分类中，I2PEMD 作为对齐度量来对未标记数据进行排序和精选，以实现更准确和稳健的伪标记。全面实验的结果证明了该方法的有效性。
