Masters-Thesis-UNet-repository

The purpose of this repo is to detail the code utilise for the U-NET segmentation used for the CHAOS dataset

There are three notebooks utilised in this repository

1. multiclass_segmentation_(camvid).ipynb: This is the main training notebook which sets up each model for training and testing.

2.Testing_dataset_unet.ipynb: This suite in jupyter notebook was used for evaluating performance of each model against one another in testing.

3.Plotting neural network results (1).ipynb: This notebook is used to produces per epoch visualisations for the training and testing completed.

Requirements

Please note that segmentation_model api was used for this analysis as ahigher level version of keras as a result this needs to be downloaded and implemented for any of the jupyter notebooks to operate effectively.