About This Project
Overview
A TensorFlow computer-vision pipeline trained on labeled X-ray data to classify bone fractures, with OpenCV preprocessing and Matplotlib-driven evaluation metrics to validate model performance.
Motivation
I wanted hands-on experience applying computer vision to a real diagnostic problem instead of another generic image classifier, so I picked X-ray fracture detection.
Lessons Learned
Preprocessing quality mattered enormously — inconsistent X-ray contrast and cropping hurt accuracy more than any architecture change I tried, which pushed me to invest more time in OpenCV preprocessing than model tuning.