HipMRI Study Segmentation with an Improved 2D UNet (Task 3) | s4744546#273

Open

baulch-m wants to merge 27 commits intoshakes76:topic-recognitionfrom

baulch-m:topic-recognition

Commits on Nov 3, 2025

First commit with readme and required python files
baulch-m
committed
added a working data loader for the OASIS dataset
baulch-m
committed
added a rough convolution component for unet model
baulch-m
committed
Updated readme and added gitignore
baulch-m
committed
re-formatted modules.py to include encoding and decoding and added a dice loss class, and added constants to train.py
baulch-m
committed
added a trainer for the unet model
baulch-m
committed
added hipmri study to gitignore
baulch-m
committed
changing task to using 2D UNet for HipMRI Prostate Cancer study - made
baulch-m
committed
added segmentation masks to the loader
baulch-m
committed
converted hipmri data into pytorch dataloaders for compatibility
baulch-m
committed
cached loaded datasets and changed train.py to work ith hipmroo dataset instead
baulch-m
committed
deleted diceloss class, added residual blocks in module.py
baulch-m
committed
reconfigured dataset.py to account for HIPMRI folder, deleted old OASIS dataset.py
baulch-m
committed
added plotting to train.py and reconfigured for dice loss function
baulch-m
committed
removed dice loss calculations and obsolete DoubleConv class
baulch-m
committed
added prediction functionality (model evaluation)
baulch-m
committed
removed DoubleConv class for real this time
baulch-m
committed
updated readme and output images (training and dice loss and best model)
baulch-m
committed
added some output predictions to /outputs for visualisation examples
baulch-m
committed
Added functionality for prediction visualisation
baulch-m
committed
Refactored some class names in dataset.py
baulch-m
committed
updated README
baulch-m
committed
finalised module and training architectures
baulch-m
committed
updated README
baulch-m
committed
updated gitignore
baulch-m
committed
test commit to get pr to work
baulch-m
committed
dummy commit
baulch-m
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

HipMRI Study Segmentation with an Improved 2D UNet (Task 3) | s4744546#273

HipMRI Study Segmentation with an Improved 2D UNet (Task 3) | s4744546#273
baulch-m wants to merge 27 commits intoshakes76:topic-recognitionfrom
baulch-m:topic-recognition

Commits on Nov 3, 2025

First commit with readme and required python files

added a working data loader for the OASIS dataset

added a rough convolution component for unet model

Updated readme and added gitignore

re-formatted modules.py to include encoding and decoding and added a dice loss class, and added constants to train.py

added a trainer for the unet model

added hipmri study to gitignore

changing task to using 2D UNet for HipMRI Prostate Cancer study - made

added segmentation masks to the loader

converted hipmri data into pytorch dataloaders for compatibility

cached loaded datasets and changed train.py to work ith hipmroo dataset instead

deleted diceloss class, added residual blocks in module.py

reconfigured dataset.py to account for HIPMRI folder, deleted old OASIS dataset.py

added plotting to train.py and reconfigured for dice loss function

removed dice loss calculations and obsolete DoubleConv class

added prediction functionality (model evaluation)

removed DoubleConv class for real this time

updated readme and output images (training and dice loss and best model)

added some output predictions to /outputs for visualisation examples

Added functionality for prediction visualisation

Refactored some class names in dataset.py

updated README

finalised module and training architectures

updated README

updated gitignore

test commit to get pr to work

dummy commit