基于深度学习的医学图像处理研究进展

Research Progress of Medical Image Processing based on Deep Learning

基于深度学习的医学图像处理已成为该领域研究的热点。深度学习方法在各种医学图像应用中取得了优异性能,达到甚至超过了专家级医生的水平。本文首先简述深度学习模型的基本原理,尤其是监督学习算法中的各种神经网络,然后总结它们在医学图像分类与识别、定位与检测、分割、配准与融合等应用领域的研究进展,最后探讨医学图像处理深度学习方法面临的挑战及应对措施。

Medical image processing based on deep learning has become a hot research topic in the academia. Deep learning algorithms achieved excellent performance in various medical image applications, reached or even exceeded the level of expert doctors. In this paper, the basic principles of the deep learning model, especially various neural networks in supervised learning algorithms, are briefly introduced. Then, the research developments are summarized in the application fields, such as medical image classification and recognition, positioning and detection, segmentation, registration and fusion. Finally, the challenges of the deep learning method for medical image processing and response measures are discussed.