Scholarlink

Artifacts in MRI images are major factors that cause image blurring and interfere with diagnosis. Advanced deep learning models are rapidly evolving and play a crucial role in smart healthcare. In particular, they improve the quality of MRI images corrupted by artifacts. However, only a few studies have systematically analyzed model architectures and usage methods for artifact correction. This study first proposes evaluation methods for correction performance, including both subjective and objective assessments. It then reviews recent progress in six types of state-of-the-art deep learning correction algorithms. The review highlights design strategies, structural innovations, and key aspects for future model development. Finally, the study discusses current technical challenges and outlines research directions for future advancements.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Survey of Deep Learning Models from DCNNs to Transformers for Correcting MRI Artifacts and Assisting Smart Healthcare

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Dingxi Liu,
  • Lichuan Ning,
  • Yuanmin Xie

摘要

Artifacts in MRI images are major factors that cause image blurring and interfere with diagnosis. Advanced deep learning models are rapidly evolving and play a crucial role in smart healthcare. In particular, they improve the quality of MRI images corrupted by artifacts. However, only a few studies have systematically analyzed model architectures and usage methods for artifact correction. This study first proposes evaluation methods for correction performance, including both subjective and objective assessments. It then reviews recent progress in six types of state-of-the-art deep learning correction algorithms. The review highlights design strategies, structural innovations, and key aspects for future model development. Finally, the study discusses current technical challenges and outlines research directions for future advancements.