Computer vision · Explainable AI
XAI Responsible AI Lung Disease Scanner
A Responsible AI research system for detecting 14 pulmonary conditions from chest X-ray images, with Grad-CAM visualisation alongside predictions.
TensorFlowKerasOpenCVGrad-CAMFlaskPython
The context
A prediction alone does not explain which image regions influenced a model. This project explores clinician-interpretable visual explanations for chest X-ray analysis.
The engineering
- Developed the detection system using TensorFlow and Keras.
- Added Grad-CAM visualisations to surface image regions relevant to the model's predictions.
- Built the application with Flask, OpenCV, Python, and Docker.
The outcome
- Detection across 14 pulmonary conditions
- Visual explanations alongside diagnostic predictions
- A research project in responsible medical AI
Next projectResponsible AI Face Authentication System
↗