3D super resolution micro CT for ice cores

 This dataset consists of two sets of micro CT scans of ice core samples with different densities. One set with high resolution (30 microns) for generating ground truth and another set of scans with low resolution (120 microns) as input. Both of these scans have been scaled to 60 microns with (downsampling or upsampling (interpolation). Then these two sets are registered with SimpleITK to create a rich dataset for 3D micro CT super-resolution tasks for Artificial intelligence models. The dataset could be used for training/testing different deep learning models for 3D resolution enhancement of computer tomography data. The relevant paper soon will appear in the author's ResearchGate.

本数据集包含两组不同密度的冰芯样本显微CT扫描数据：一组为30微米的高分辨率扫描结果，用于生成基准真值；另一组为120微米的低分辨率扫描结果，作为模型输入。两组扫描数据均已通过下采样或上采样（插值）操作统一缩放至60微米分辨率。随后，使用SimpleITK对这两组数据进行配准，从而构建出适用于人工智能模型的3D显微CT超分辨率任务专用数据集。本数据集可用于训练、测试各类深度学习模型，以实现计算机断层扫描数据的3D分辨率增强。相关研究论文即将在作者的ResearchGate平台发布。