Fracture Segmentation Application on Coal CT images

Published: 5 November 2023| Version 1 | DOI: 10.17632/krg7wnxmzk.1
shanilka fernando


This application can be utilized for segmenting fractures in coal CT images. You have the option to employ three distinct deep learning architectures for segmenting the coal CT images.


Steps to reproduce

The Brazilian disc test was conducted on coal specimens, and CT images were subsequently acquired. Following this, fractures within the CT images were segmented. The resulting segmented images were then used to create masked images for training three deep learning architectures (UNet, UNet transfer, and DeepLabV3+). Once the deep learning models were trained, they were saved in h5 format. Subsequently, a desktop application was developed using PyCharm and PYQT5 designer.


Shandong University of Science and Technology


Image Segmentation, Mining Engineering, Deep Learning