Skin Cancer Lesion Classification using HAM10000 Dataset

This is a Skin Cancer Lesion Classification project.In this repository, I tackle the task of classifying dermatoscopic images into different categories of skin cancer lesions using the HAM10000 dataset. Skin cancer is a significant health concern, and early detection through image analysis can be a powerful tool in aiding diagnosis.

Dataset Overview

The HAM10000 dataset Kaggle is a collection of dermatoscopic images of skin lesions, containing seven classes of skin cancer lesions:

Skin Cancer Lesion Classes

Class	Description
Melanocytic nevi (nv)	Melanocytic nevi, also known as moles, are benign skin lesions consisting of melanocytes.
Melanoma (mel)	Melanoma is a malignant skin cancer that arises from melanocytes. It is the most dangerous form of skin cancer.
Benign keratosis-like lesions (bkl)	Benign keratosis-like lesions include various non-cancerous skin conditions that resemble actinic keratoses or basal cell carcinomas.
Basal cell carcinoma (bcc)	Basal cell carcinoma is a common form of skin cancer that arises from basal cells in the epidermis.
Actinic keratoses (akiec)	Actinic keratoses, also known as solar keratoses, are precancerous lesions caused by sun exposure.
Vascular lesions (vas)	Vascular lesions include various blood vessel-related skin conditions, such as angiomas.
Dermatofibroma (df)	Dermatofibroma is a benign skin condition characterized by fibrous tissue growth in the dermis.

Project Structure

• data/: Placeholder for the dataset (not included in the repository).
• models/: Trained machine learning models.
• src/: Source code for data preprocessing, model training, and evaluation. All Jupyter files.
• requirements.txt: List of project dependencies.
• README.md: This file.

Getting Started

1. Installing Dependencies:

   pip install -r requirements.txt

2. Download and Prepare Dataset:

   • Download the HAM10000 dataset from Kaggle.
   • Extract the dataset files into the data/ directory.
   • Run data preprocessing scripts in the src/ directory to prepare the dataset.

4. Exploration and Model Training:

   Explore the Jupyter notebooks in the src/ directory for data analysis and model training.

5. Evaluate Models:

   You can evaluate the trained models using the provided scripts in the src/ directory.

Acknowledgments

• HAM10000 Dataset
• Original Research paper
• Digital Sreeni

Thank you for your interest in our project!