The goal of this Master's thesis is to train and optimise a Convolutional Neural Network to identify the atrium of the heart from CT scan images of the chest.
This folder is organised in different sections
- code: All the code used to generate the data, train the network, conduct experiments, and process the eperimental results.
- ct_atrium: The CT scans used to generate the datasets. These are not uploaded unto Github as the information is private and sensitive.
- datasets: Directory where all the generated datasets are located. These are generated locally using scripts stored in ../code/dataset_generation and not uploaded unto Github due to the size of the resulting HDF5 files.
- experimental results: Where all the experimental results are stored in folders with self-explanatory names indicating the nature of the experiments conducted.
- Thesis Report: Final master's thesis submitted. The final report isn't has nice and polished as it could have been but it was written amid a severe bout of sciatica which left me in paralysed with pain for 2 months in the middle of the summer.