This repository contains the code and documentation for a machine learning project focused on classifying melanoma cancer images. The project leverages a dataset of meticulously curated images to develop a model capable of distinguishing between benign and malignant skin lesions.
Melanoma is one of the most dangerous forms of skin cancer, known for its rapid spread and high mortality rate if not diagnosed early. The early detection of melanoma can significantly improve survival rates. This project aims to harness the power of deep learning to aid in the early diagnosis of melanoma by classifying images of skin lesions.
The dataset used in this project consists of 13,900 images of skin lesions, each uniformly resized to 224 x 224 pixels.
The primary objective of this project is to develop a reliable and accurate machine learning model that can classify skin lesion images into benign or malignant categories. The model should assist dermatologists and healthcare professionals in making faster and more accurate diagnoses, ultimately contributing to better patient outcomes.
This project employs a variety of machine learning and deep learning techniques to achieve its goals. The analysis follows a structured approach, outlined below:
1. Data Preprocessing
- Resizing images to 224 x 224 pixels.
- Data augmentation techniques such as rotation, horizontal flipping, zooming, shearing, dan fill mode to enhance model robustness.
2. Model Architecture
Convolutional Neural Networks (CNNs): The core model is based on CNN architectures due to their proven effectiveness in image classification tasks.
3. Training and Optimization
- Loss Function: The model is compiled with binary cross-entropy as the loss function, which is suitable for the binary classification task of distinguishing between benign and malignant skin lesions.
- Optimizer: The Adam optimizer is utilized for training, offering efficient and adaptive learning capabilities to improve model performance.
4. Evaluation Metrics
- Accuracy: Overall accuracy of the model in correctly classifying the images.
- Precision, Recall, and F1-Score: To assess the balance between sensitivity and specificity.
- Confusion Matrix: To provide detailed insights into the classification results, including false positives and false negatives.
5. Model Deployment
The final model, after validation and testing, is deployed using Hugging Face.
- Analysis and Visualization (EDA)
- Feature Engineering
- Computer Vision
- Deep Learning
- Data Scientist