s4986525 Siamese Submission

[IsaiahWu]

This pull request contains the full working implementation of the Siamese neural network model designed to classify skin lesions as benign or melanoma reaching an accuracy of 80%
The project includes:

• Model code with three convolutional layers and contrastive loss
• Dataset preparation scripts including augmentation for class balance
• Training and validation scripts with Hyperparameters optimized for accuracy
• Evaluation metrics scripts generating confusion matrix, ROC curves, and t-SNE plots
• A README documenting installation, usage, dataset preprocessing, and results

Incorporation of Feedback

• The model show a high accuracy due to data imbalance

Description and Comments:

• Each script file contains comments, explaining the purpose of functions and critical code sections.
• The README file provides background on the problem, dataset, model details, training process, evaluation methods, and future improvement plans.

The pull request message itself summarizes all relevant components to guide reviewers.