Deep learning assisted processing of synchrotron-based micro-CT imaging of fast, dynamic multiphase flow in porous media

These are the datasets used for training, testing and validation of an implementation of the U-ResNet model published by Tang et al. (2022) used for multiphase segmentation of porous media. In this work, the above mentioned model was used to segment synchrotron-based micro-CT scans of water-wet and mixed-wet Bentheimer sandstone sample acquired during drainage coreflooding experiments with a temporal resolution of 1s with 15-16s time lapse intervals. The uploaded datasets are as follows, note that XX represent the different slice numbers used for training / testing: 1) MW_train_XX_gray.tif = 2D gray scale slices used for training the model to segment mixed-wet sample. 2) MW_train_XX_seg.tif = Corresponding segmented images of the 2D gray scale slices used for training the model to segment the mixed-wet sample. 3) MW_test_XX_gray.tif = 2D gray scale slices used for testing the trained model to segment mixed-wet sample 4) MW_test_XX_seg.tif = Corresponding segmented images of the 2D gray scale slices used for testing the trained model to segment the mixed-wet sample. 5) WW_train_XX_gray.tif = 2D gray scale slices used for training the model to segment water-wet sample 6) WW_train_XX_seg.tif = Corresponding segmented images of the 2D gray scale slices used for training the model to segment the water-wet sample. 7) WW_test_XX_gray.tif = 2D gray scale slices used for testing the trained model to segment water-wet sample 8) WW_test_XX_seg.tif = Corresponding segmented images of the 2D gray scale slices used for testing the trained model to segment the water-wet sample. 9) Inference_Mixed-wet_drainage_fb21.tif = 3D tif image of size 1536x1536x160 voxels that was used for inference / segmenting the mixed-wet sample using the trained and tested model. 10) Inference_Water-wet_drainage_fb30.tif = 3D tif image of size 1536x1536x160 voxels that was used for inference / segmenting the water-wet sample using the trained and tested model.