医学图像伪造检测技术研究综述

A Review of Research on Medical Image Forgery Detection Technology

摘要: 医学图像伪造检测技术是保障医疗数据安全、确保临床诊断准确性的重要研究方向。随着数字图像处理技术和生成式人工智能的快速发展，医学图像伪造手段日益复杂，其检测技术也在不断完善。近年来，医学图像伪造检测技术的突破性研究日新月异，为了全面了解该领域的研究进展，本文对医学图像伪造检测的核心问题进行系统综述：首先，分类总结伪造检测方法及其演进历程；其次，详细介绍了常用数据集的特点与适用范围以及评估指标；重点梳理了医学CT、MRI图像小区域伪造和抗压缩干扰的优化策略；最后，提出医学图像伪造检测技术研究的未来方向，是深度学习的多模态融合技术和主动防御策略的医学信息伪造检测研究，以应对生成式AI时代的医学图像伪造挑战。

Abstract: Medical image forgery detection technology is an important research direction to ensure the safety of medical data and the accuracy of clinical diagnosis. With the rapid development of digital image processing technology and generative Artificial intelligence, the means of medical image forgery are increasingly complex, and its detection technology is also constantly improving. In recent years, the breakthrough research of medical image forgery detection technology is changing rapidly. In order to fully understand the research progress in this field, this paper systematically reviews the core issues of medical image forgery detection: first, classify and summarize forgery detection methods and their evolution process; Secondly, the characteristics, application scope and evaluation indicators of common data sets are introduced in detail; Focused on the optimization strategy of medical CT and MRI image small area forgery and anti-compression interference; Finally, the future direction of research on medical image forgery detection technology is proposed, which is the multimodal fusion technology of deep learning and the research on medical information forgery detection of active defense strategy, to meet the challenge of medical image forgery in the era of generative AI. This electronic document defines the standard format of the Chinese academic journals published by the Hans Publishing. The elements such as the paper title, author, affiliation, abstract, section title, main text, figure, table and references are defined, and this document is formatted according to the Hans standard, which illustrates all the formats.