Cervical Cancer Detection Using Deep Learning

Medical image classification using efficientNetB0 achieved over 97.7% accuracy.

Enhanced model generalization by employing more image preprocessing techniques such as image resizing, normalization, and augmentation.

Addressed image imbalance and enhanced the diversity of the dataset by implementing data augmentation methods.

Presented confusion matrices and generated accuracy/loss graphs to visualize how well the model has been trained.

Incorporated steps in the process of glaucoma diagnosis to lessen the burden on medical specialists and improve early disease detection.

The project contributes to the growth of AI in medical imaging.

Model Comparisons between which one is the best modeled multiple models (CNN, EfficientNet, Sequential).

Identified strengths and areas where performance needs to be improved with a visualized lens.

We used pre-trained weights to do training in less time and still get the performance.