The project aims to develop a deep learning model for accurate pneumonia detection from chest X-ray images. This technology can enhance diagnostic speed and accuracy, benefiting healthcare professionals, patients, and healthcare systems.
Develop and fine-tune a deep learning model for pneumonia detection. Apply data augmentation and optimization techniques. Evaluate model performance using metrics like accuracy, recall, and loss.
Dataset: 5,863 chest X-ray images categorized as "Pneumonia" and "Normal". Origin: Pediatric patients aged 1-5 years from clinical care at Guangzhou Women and Children’s Medical Center.
Baseline Model Dense Neural Network achieved a training accuracy of 93.94% and validation accuracy of 87.50%. CNN Models Model 2: Basic CNN showed overfitting with a training accuracy of 80.41% and validation accuracy of 56.25%. Model 3: Enhanced CNN architecture achieved a training accuracy of 98.66% and validation accuracy of 75.00%.
Model 3 outperformed Model 2 with better metrics: lower test loss, higher accuracy, and precision.
I deployed Model 3 using flask and it was able to make the predictions
