Permeability Prediction in 2D: Dataset and Trained Convolutional Neural Networks
Published in Mendeley Data, 2024
The dataset linked contains the following:
- Original microstructure data, 4500 total images
- External test set, 400 total images
- Permeability for the entire set
- Pressure and velocity maps for the entire set
- Trained CNN models for permeability prediction:
- AlexNet
- DenseNet
- ResNet34
- ResNet50
- VGG19
- VGG19 (top 10 Ensemble)
Recommended citation: Adam, Andre; Stallard, Silven; Fang, Huazhen; Li, Xianglin (2024), “Permeability Prediction in 2D: Dataset and Trained Convolutional Neural Networks”, Mendeley Data, V1, doi: 10.17632/576dvrrsdx.1
Download Paper