Abstract: Bone age, one of the important basis for age assessment, is a very active research field in forensic medicine. The current bone age assessment （BAA） methods are mainly based on the manual assessment of the gradual anatomical changes in X-ray images, which are time-consuming and prone to inter and intra-reviewer variability. Moreover, it is suffered ethical debate that the radioactive X-ray method is for inspection instead of treatment. Magnetic resonance imaging （MRI） is the preferred method for BAA without radiation. The MRI system produces bone image as three-dimensional volumetric data in the form of a stack of two-dimensional slices. MRI has a superior ability in providing the images of cartilage of epiphyseal plate, and can clearly show the growth and development of epiphysis. The use of skeletal MRI for BAA has become a feasible strategy to reduce radiation risks. At the same time, with the development of deep learning technology, its advantages in image analysis and big data processing are emerging. Therefore, the combination of deep learning and MRI for BAA has been the research frontier in recent years. This paper reviewed the research progress on BAA using deep learning and MRI, and summarized the new technologies and methods for BAA, in order to provide new ideas and reference for future works.

Key words: forensic anthropology, age assessment by skeleton, deep learning, magnetic resonance imaging （MRI）

摘要： 骨龄作为年龄认定的重要依据之一，是法医学的重点研究内容。现有主流骨龄评估方法多基于骨关节X线片，虽已达到较高的准确性，但此类方法主要基于人工阅片，耗时、依赖经验，且非诊疗目的的放射性检查当下备受争议。磁共振成像（magnetic resonance imaging，MRI）是无辐射的三维断层图像，具有优越的骺板软骨成像能力，可清晰显示骨骺生长发育情况，将骨骼MRI用于骨龄评估已成为解决辐射风险的可行策略。同时，随着深度学习的不断进步与发展，使其在图像分析和大数据处理方面的优势逐渐显现。因此，将深度学习与骨骼MRI有机融合进行骨龄评估已成为近年来法医骨龄评估领域新的研究前沿和热点。基于此，通过综述近年来国内外基于深度学习的MRI骨龄评估研究，系统总结了基于深度学习及骨骼MRI的骨龄评估新技术、新方法，以期为骨龄评估研究提供新的思路与参考。

关键词: 法医人类学, 骨龄评估, 深度学习, 磁共振成像