2023-24

Melanocytic nevi classification

A melanocytic nevi classification system using Transfer Learning to assist in early detection of skin cancer.

About the Project

This project utilizes Transfer Learning with Resnet50 to classify melanocytic nevi, aiding in the early detection of skin cancer. The system is designed to improve diagnostic accuracy and support medical professionals.

Key Features

• →Automated Skin Disease Classification
• →Transfer Learning with ResNet50
• →Preprocessing & Data Augmentation
• →Binary Cross-Entropy Optimization
• →Accurate model performance

Model architecture

What I Learned

Preprocessed the melanocytic nevi dataset, implemented data augmentation, tuned hyperparameters for optimal model performance, tested model on unseen data and deployed using Flask.

Model Performance

The confusion matrix provides an in-depth evaluation of the classification model's performance: True Positives (Melanocytic Nevi): 1195 samples correctly identified. True Negatives (Normal Skin): 586 samples correctly classified. False Positives: Only 1 normal skin sample misclassified as melanocytic nevi. False Negatives: 0 cases; the model consistently identified all melanocytic nevi. False Negatives: 0 cases; the model consistently identified all melanocytic nevi.

Published Paper