使用机器学习、深度学习方法进行医学图像去噪：综述。
Denoising Medical Images Using Machine Learning, Deep Learning Approaches: A Survey.
机构信息
Department of Electrical Engineering, Central Tehran Branch, Islamic Azad University, Tehran, Iran.
Department of Electrical Engineering, Majlesi Branch, Islamic Azad University, Isfahan, Iran.
出版信息
Curr Med Imaging. 2021;17(5):578-594. doi: 10.2174/1573405616666201118122908.
OBJECTIVE
Several denoising methods for medical images have been applied, such as Wavelet Transform, CNN, linear and Non-linear methods.
METHODS
In this paper, A median filter algorithm will be modified and the image denoising method to wavelet transform and Non-local means (NLM), deep convolutional neural network (Dn- CNN), Gaussian noise, and Salt and pepper noise used in the medical image is explained.
RESULTS
PSNR values of the CNN method are higher and showed better results than different filters (Adaptive Wiener filter, Median filter, and Adaptive Median filter, Wiener filter).
CONCLUSION
Denoising methods performance with indices SSIM, PSNR, and MSE have been tested, and the results of simulation image denoising are also presented in this article.
目的
已经应用了几种医学图像去噪方法，如小波变换、CNN、线性和非线性方法。
方法
本文将对中值滤波器算法进行修改，并解释用于医学图像的小波变换和非局部均值（NLM）、深度卷积神经网络（Dn-CNN）、高斯噪声和椒盐噪声的图像去噪方法。
结果
CNN 方法的 PSNR 值更高，并且比不同的滤波器（自适应维纳滤波器、中值滤波器和自适应中值滤波器、维纳滤波器）显示出更好的结果。
结论
使用 SSIM、PSNR 和 MSE 指标对去噪方法的性能进行了测试，并在本文中还呈现了模拟图像去噪的结果。
Zotero 插件