🧠 Intracranial Hemorrhage Detection using Deep Learning

In this project we detect and diagnose the type of hemorrhage in CT scans using Deep Learning!

The training code is available in train.py. Code for the metrics reported in the paper is available in notebooks/Week 11 - tlewicki - metrics clean.ipynb.

Setup instructions:

1. The dataset can be downloaded from here (the dataset 500 GB). There's a dowloaded and unzipped version on SJSU HPC disk at: '/data/cmpe257-02-fa2019/team-1-meerkats/rsna-intracranial-hemorrhage-detection/'

2. Install the requirements and activate the virtual environment:

conda env create -f environment.yml
conda activate brainenv

3. Install pydicom (not available from conda distribution)

(brainenv) pip install pydicom

4. Download the weights file and place it in ichdemo/ folder

5. Navigate to ichdemo/ & and run the demo app

(brainenv) python3 app.py

6. Navigate your browser to http://localhost:9999/ to interact with the demo app

7. You can run any notebooks in notebooks/ dir by typing:

(brainenv) jupyter notebook