Project Status: Finished

Project Intro/Objective

This project, conducted at the University of Applied Sciences and Arts Northwestern Switzerland FHNW, School of Engineering, in Brugg-Windisch, represents a pioneering effort in utilizing machine learning (ML) and deep learning (DL) techniques for the early detection of prolactinomas. Spearheaded by Jan Zwicky, Thomas Mandelz, and guided by Professors Suter and Perruchoud, the study collaborates closely with medical professionals from Kantonsspital Aarau, Dr. Tristan Struja and Dr. Felice Burn.

Folderoverview

Folder	Subfolders	Description
data	test, train	Data for testing and training
eda		Exploratory Data Analysis (EDA) files, Data processing and partitioning files
modelling	combined, mri_data, tabular_data	Modelling-related folders: combined models, MRI data, tabular data
	- mri_data: results_augmented_weighted	MRI data and augmented weighted results
	- tabular_data:	tabular data machine learning pipelines
	- images:	Images for tabular data analysis: fn, fp, tn, tp
	- results_csv	CSV results for tabular data analysis
models		Saved machine learning models and related components
	- saved_models	Pretrained or fine-tuned models saved for future use
	- MedicalNet	MedicalNet (Med3d) library components for medical imaging
	- NODE	NODE (Neural Oblivious Decision Ensembles) for deep learning
raw_data	nii_files	Raw data for tab and NII files
src		Source code for deep learning pipeline and related files

Methods Used

• Deep Learning
• Machine Learning

Technologies

• Python
• PyTorch
• wandb
• numpy
• pandas
• Azure Machine Learning

Getting Started

• Clone this repo (for help see this tutorial).
• Demo files are being kept here
• Raw Data is being kept here
• Explorative Dataanalysis Scripts and Files are being kept here
• Megadetector Scripts and data is being kept here
• Models are being kept here
• Models submissions are being kept here
• Source files for training are being kept here
• Source files for pipeline are being kept here

Pipenv for Virtual Environment

First install of Environment

• open cmd
• cd /your/local/github/repofolder/
• pipenv install
• Restart VS Code
• Choose the newly created "tierli_ahluege" Virtual Environment python Interpreter

Environment already installed (Update dependecies)

• open cmd
• cd /your/local/github/repofolder/
• pipenv sync

Contributing Members

• Thomas Mandelz
• Jan Zwicky