Skin Cancer Detection

Developed a deep learning model using EfficientNetB0 to classify skin cancer images from a Kaggle

dataset of 10,000 images, achieving 92% test accuracy. Preprocessed images with OpenCV and ImageDataGenerator, applying data augmentation and

normalization, increasing model robustness by 15% against overfitting. Evaluated model performance using confusion matrix and classification report, reducing false positives by

20% through hyperparameter tuning. Visualized results with Matplotlib, presenting findings to 30+ stakeholders, resulting in 85% approval for

model deployment potential.