PNEUMONIA DETECTION USING MOBILENETV2 | Python, Tensorflow, Keras, MobileNetV2, CNN, Deep learning

• Enhanced the technique for pneumonia detection, achieving a reliable 91% accuracy in identifying pneumonia cases and an impressive 90% precision for recognizing normal cases

• Effectively addressed the dataset’s class imbalance, by combining an ensemble approach with under-sampling, ensuring good precision for both Normal and Pneumonia Cases.

Distribution of Train Dataset

Distribution of Test Dataset

Normal chest X-Rays

Pneumonia Chest X-Rays

Accuracy Score

Classification Report(Without applying SMOTE)

Classification Report(SMOTE APPLIED)

Confusion matrix

Accuracy Report