SKIN CANCER PREDICTION

OVERVIEW

A computer-vision image classification project applying convolutional neural networks to classify benchmark skin image datasets. This project is an engineering and machine learning experiment — not a medical diagnostic tool.

THE IDEA

Apply CNN-based image classification to an image dataset as a practical deep learning study in image preprocessing, convolutional architecture design, training, and test set evaluation.

WORKFLOW

KEY FEATURES

• Image input and normalization pipeline
• Convolutional neural network architecture
• Training and validation workflow
• Multi-class classification output
• Model evaluation on benchmark test data

ENGINEERING APPROACH

The pipeline applies image resizing, normalization, and standard augmentation prior to training. A CNN architecture is trained and evaluated on the classification task. Note: this is an academic computer-vision image classification experiment and not a medical diagnostic tool.