Bone & Bone Lesion Segmentation in SPECT/CT Using Res-U-Net
This is almost the same as the U-Net segmentation project (SPECT-CT-Seg-UNet), but the U-Net was modified to the Residual U-Net discribed in:
Zhange, Zhengxin, et al. "Road extraction by deep residual u-net." IEEE Geoscience and Remote Sensing Letters, vol. 15, no. 5, pp. 749-753, May 2018.
Network was trained with the SPECT/Attuenuation-Map simulations and tested with the patient PET/CT
Main script: run/train_ResUNet.py
Example Segmentation Results
Validation Set (SPECT/CT simulations)
Dice similarity coefficient: 0.977 for lesion, 0.979 for bone
Resulting an undesired result with the unseen patient PET/CT images.
