Abstract: Fetal brain MRI, due to its non-invasiveness, absence of radiation, and high soft-tissue contrast, has become an important tool for assessing fetal brain development and diagnosing congenital brain abnormalities. High-quality fetal brain MR images play an important role in clinical diagnosis, treatment, and scientific research of fetal brain development. Image processing techniques can enhance the quality of fetal brain MRI images, meeting the requirements for diagnosis and research, thus the studies in this field hold significant importance. This paper provides a brief introduction to fetal brain structure and its MR image datasets, and elaborates on six techniques, including image quality assessment, image registration, image denoising, image bias field correction, image artifact correction, and super-resolution reconstruction. Firstly, the importance of image processing technologies for fetal brain MR images was presented. Subsequently, the structure of the fetal brain and its MRI datasets were introduced. Then the six image processing techniques are introduced respectively. The research status both at home and abroad is systematically described. The performance of different methods is compared and analyzed. And the current achievements and challenges are summarized respectively. Finally, the existing issues and future research directions in the field of fetal brain MR image preprocessing are discussed from the perspectives of technology and clinical application.

Key words: fetal brain, MR image, image processing, deep learning

摘要： 胎儿脑核磁共振成像技术因其无创、无辐射和高软组织对比度，已成为评估胎儿大脑发育和诊断先天性脑异常的重要工具。高质量的胎儿脑核磁共振图像在临床诊疗和胎儿脑发育等科学研究方面发挥着重要作用。图像处理技术可提升胎儿脑核磁共振图像质量，满足诊断与研究需求，故其在胎儿脑核磁共振图像领域的研究具有重要意义。本文对胎儿脑结构及其核磁共振图像数据集进行简要介绍，并对图像质量评价、图像配准、图像去噪、图像偏差场校正、图像去伪影及超分辨率重建六个方面的技术进行阐述。首先阐述了图像处理技术应用于胎儿脑核磁共振图像的重要性；随后介绍了胎儿脑结构及其核磁共振图像数据集；接着分别就六个方面的图像处理技术进行详细介绍，系统地阐述了不同技术的国内外研究现状，对不同方法的性能进行比较与分析，并对已取得的成果与面临的挑战进行了小结；最后从技术、临床应用等角度探讨了胎儿脑核磁共振图像处理领域存在的问题和未来的研究方向。

关键词: 胎儿脑, 核磁共振图像, 图像处理, 深度学习