LSRML: A latent space regularization based meta-learning framework for MR image segmentation
作者：
Highlights：
• This work incorporates domain discriminator and image reconstruction into a meta-learning framework for prostate MRI segmentation, and achieves better segmentation and generalization performance than state-of-the-art methods.
• This work constructs a large-scale multimodal prostate dataset, which has significant advantages in terms of the quantity of data, data type, and data annotation.
• This work analyzes the impact of domain shift on deep learning models through extensive experiments on the proposed and the benchmark datasets.
摘要
•This work incorporates domain discriminator and image reconstruction into a meta-learning framework for prostate MRI segmentation, and achieves better segmentation and generalization performance than state-of-the-art methods.•This work constructs a large-scale multimodal prostate dataset, which has significant advantages in terms of the quantity of data, data type, and data annotation.•This work analyzes the impact of domain shift on deep learning models through extensive experiments on the proposed and the benchmark datasets.
论文关键词：Latent space regularization,Meta learning,Domain generalization,Domain discriminator,Multi-source domain adaptation
论文评审过程：Received 18 January 2022, Revised 7 April 2022, Accepted 30 May 2022, Available online 31 May 2022, Version of Record 5 June 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108821
