| 摘要: |
| ［摘要］ 近年来，日渐成熟的人工智能深度学习技术使得众多领域逐渐实现自动化智能化作业。在医疗领域，随着医疗数据电子化和互联网医疗的发展，基于卷积神经网络实现包含定位、分割和分类于一体的辅助诊断系统应用已成为新型医疗模式发展的必然趋势。医学影像分割技术是医疗图像自动分析中的难点和重点，目前仍面临许多亟待解决的问题。该文将从临床医学影像的特点、深度学习主流分割网络和医学图像分割网络在临床中的应用3个方面对医学图像分割领域的研究进展进行系统综述，并进一步分析卷积神经网络在医学影像分割任务中的发展现状、面临的挑战以及未来的发展方向。 |
| 关键词: 深度学习 医学影像 卷积神经网络 图像分割算法 |
| DOI：10.3969/j.issn.1674-3806.2020.02.02 |
| 分类号:TP 18 |
| 基金项目:国家自然科学基金项目（编号：61603182）；中国博士后科学基金项目（编号：2019M661832）；江苏省博士后科研资助计划项目（编号：2019K226） |
|
| Medical image segmentation methods based on deep learning |
|
YOU Qi-jing, WAN Cheng
|
|
College of Electronic and Information Engineering, Nanjing University of Aeronautics and Astronautics, Jiangsu 211106, China
|
| Abstract: |
| ［Abstract］ In recent years, the increasingly developed technology of deep learning of artificial intelligence makes many fields gradually realize automatic intelligent work. In the field of medicine, with the developments of medical data electronization and internet medicine, it has become an inevitable trend to develop a new medical mode to realize computer-aided diagnosis systems based on convolutional neural networks, which includes positioning, segmentation and classification. Medical image segmentation technology is the difficulty and key point in the automatic analysis of medical image. At present, there are still many problems to be solved. In this paper, the progress of medical image segmentation will be systematically reviewed from three aspects: the characteristics of clinical medical image, the introduction of deep learning mainstream segmentation networks and the application of current medical image segmentation networks in clinical application, and the current development situation, challenges and future development direction of convolution neural networks in medical image segmentation task will also be analyzed. |
| Key words: Deep learning Medical image Convolution neural networks Image segmentation |
