CardioVisionAI

thumbnail
data inputs
dashboard
health assessment data
home page
create account page
3d model
backside of heart

INSPIRATION Cardiovascular diseases are personal to us, as we have seen many of our own relatives suffer from them. Their experiences inspired us to explore how machine learning could help us better understand cardiovascular health. With CardioVision, we hope to turn this personal motivation into a meaningful step toward making healthcare technology more accessible and understandable.

WHAT WE BUILT CardioVision is a machine-learning application that predicts four targets: Coronary Artery Disease (CAD), Left Anterior Descending artery (LAD), Left Circumflex artery (LCX), and Right Coronary Artery (RCA). It combines prediction models with an interactive dashboard and feature-importance visualizations.

HOW WE BUILD IT We used Python, Pandas, NumPy, and Scikit-learn to process clinical data and train classification models. The project includes data exploration, prediction pipelines, and a dashboard for exploring model outputs. We also developed explainability functionality to investigate which input features influence predictions.

CHALLENGES WE FACED One of our challenges was finding a suitable, colourful 3D heart visualization that could clearly represent the different coronary arteries. We also faced challenges handling clinical datasets with different input features, ensuring patient inputs matched the model's expected schema, and presenting machine-learning predictions in an understandable way rather than just displaying results.

WHAT WE LEARNED Through this project, we explored the machine-learning workflow, from data preparation and model evaluation to prediction interfaces and model explainability. We also learned how to collaborate on a shared codebase and integrate separate project components.

FUTURE IMPROVEMENT Future work includes improving the dashboard, expanding patient-specific explanations, refining model evaluation, and exploring interactive cardiovascular visualizations.

DISCLAIMER CardioVision is an educational and research project. Its predictions are not medical diagnoses and must not replace professional medical advice.