Three-dimensional medical image super-resolution reconstruction method based on cascaded cavity convolution

The invention relates to a three-dimensional medical image super-resolution reconstruction method based on cascaded cavity convolution. The method comprises the following steps: 1, constructing a dataset and carrying out data preprocessing; 2, a three-dimensional medical image super-resolution reconstruction network based on cascaded hole convolution is built through a deep learning framework TensorFlow, and the three-dimensional medical image super-resolution reconstruction network based on cascaded hole convolution comprises three parts, namely a shallow feature extraction module, a nonlinear mapping module and a reconstruction module; 3, performing model training; and 4, inputting three-dimensional medical low-resolution small blocks, loading the model trained in the step 3, and outputting reconstructed super-resolution small blocks. 本发明涉及一种基于级联空洞卷积的三维医学图像超分辨率重建方法，包括下列步骤：第一步，构建数据集并进行数据预处理；第二步，通过深度学习框架TensorFlow搭建基于级联空洞卷积的三维医学图像超分辨率重建网络，基于级联空洞卷积的三维医学图像超分辨率重建网络包括三个部分，浅层特征提取模块、非线性映射模块和重建模块；第三步，模型训练；第四步，输入三维医学