Skin-Cancer-Diagnosis

This repository showcases a Convolutional Neural Network (CNN) module developed on Jupyter and an AutoML module implemented on Google Cloud Platform (GCP) through VertexAI. Insights include discovering a 37% probability of misclassifying Melanocytic nevi as vascular lesions by machine learning models. Moreover, both models effectively identify distinct characteristics of vascular lesions. Notably, AutoML test predictions exhibit independence from the count of images provided and demonstrate heightened efficiency in predicting BCC, MEL, and BKL skin cancer types