A common symptom of strokes is facial drooping and asym- metry. We present a Python program that aims to help patients assess their own health by analysing their face and determining the likelihood of a positive diagnosis. We pre-process and detect facial landmarks of the dataset, consisting of stroke patients and healthy individuals, with the dlib library. We then extract features from the landmarks, and finally we train an SVM model and use it to predict if the user is experiencing a stroke. The experimental results of our system average to 93%, indicat- ing a high accuracy in facial droop recognition, which help identifying patients having a stroke.
Timesheet | Slack channel | Project report |
---|
CMPT340.Facial.Droop.Project._compress.mp4
Explain briefly what files are found where
repository
├── src ## source code stroke detection program
├── dataset ## dataset that train and test the model
├── model ## pre-traind machine learning models
├── scripts ## scripts, train and save the model to model directory
├── README.md ## You are here
├── requirements.yml ## If you use conda
Provide sufficient instructions to reproduce and install your project. Provide exact versions, test on CSIL or reference workstations.
git clone $THISREPO
cd $THISREPO
conda env create -f requirements.yml
conda activate $THISREPO
To train the model
cd project_21
python scripts/train_model_svm.py
This will save the model to model directory. The model evaluation result will be printed in the terminal.
To run the GUI
cd project_21
python src/UI/UI.py
- Use git
- Do NOT use history re-editing (rebase)
- Commit messages should be informative:
- No: 'this should fix it', 'bump' commit messages
- Yes: 'Resolve invalid API call in updating X'
- Do NOT include IDE folders (.idea), or hidden files. Update your .gitignore where needed.
- Do NOT use the repository to upload data
- Use VSCode or a similarly powerful IDE
- Use Copilot for free
- Sign up for GitHub Education