Unet for segmenting Luna-16 dataset. Data analytics course. Batch Size: 3 Trained with 10 images Training with 90% Validation with 10% Tested with 20 images
Change to using softmax with 3 output Use (#labels, x, y, z, 1, 1?)
No post-processing: DICE: [] Hausdorff: Mean Distance:
DeepMedic Source Code: https://github.com/Kamnitsask/deepmedic
No post-processing: DICE: [] Hausdorff: Mean Distance:
https://journals.scholarsportal.info/pdf/02780062/v24i0008/1025_tasotplic.xml https://aapm.onlinelibrary.wiley.com/doi/epdf/10.1118/1.1312192 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4521615/
https://github.com/ellisdg/3DUnetCNN https://github.com/tensorflow/models/blob/master/samples/outreach/blogs/segmentation_blogpost/image_segmentation.ipynb https://github.com/zizhaozhang/unet-tensorflow-keras