MANGALORE UNIVERSITY
Department of PG studies and research in computer science
Mangalore University, Mangalagangotri,
Mangalore-574199
MINI PROJECT SYNOPSIS
Topic:
Skin Cancer Classification Using Machine Learning
Submitted By:
KRISHNA
3rd Semester MCA
Department of PG studies and research in computer science
Mangalore University
Submitted To:
Dr. B H Shekar
Chairman & Senior Professor
Department of PG studies and research in computer science
Mangalore University
2025-2026
1
Skin Cancer Classification Using Machine Learning
Introduction
Skin cancer is one of the most prevalent forms of cancer worldwide, and early diagnosis
plays a vital role in improving patient survival rates. Conventional diagnosis depends on
expert dermatological examination, which may be subjective and time-consuming. Machine
Learning (ML) techniques provide an effective solution by analyzing medical image features
and enabling automated classification of skin lesions. This project focuses on developing a
machine learning–based skin cancer classification system.
Problem Statement
Skin cancer is a serious disease, and early detection is very important to save lives.
Traditional skin cancer diagnosis depends on doctors, which can be time-consuming and may
vary from person to person. In some areas, expert dermatologists are not easily available.
Therefore, there is a need for a simple and automated system that can help in identifying
skin cancer at an early stage. This project aims to develop a machine learning–based system
that analyzes skin images and classifies skin cancer accurately, helping doctors in diagnosis.
Objective
The objective of this project is to design and implement a machine learning model that can
accurately classify different types of skin cancer using extracted image features, thereby
assisting healthcare professionals in early diagnosis.
Dataset Description
Dataset consists of dermoscopic skin lesion images
Collected from a public Kaggle repository
Includes cancerous (malignant) and non-cancerous (benign) images
Each image is labeled for supervised machine learning
2
Images are in JPEG/PNG format
Dataset is preprocessed (resizing, normalization)
Split into training and testing sets for evaluation
Methodology
1. Dataset Collection
The dataset consists of dermoscopic images of skin lesions along with their corresponding
class labels. These images represent different categories of skin cancer and benign lesions.
The dataset is obtained from publicly available medical image repositories.
2. Image Preprocessing
To improve image quality and ensure uniformity, several preprocessing steps are applied:
Resizing images to a fixed dimension
Noise reduction using filtering techniques
Normalization to scale pixel values
Color space conversion to enhance lesion visibility
These steps help in improving feature extraction and model performance.
3. Feature Extraction
Since traditional machine learning algorithms are used, meaningful features are extracted
from the images:
Color features such as mean and variance of RGB values
Texture features using statistical methods like GLCM
Shape features such as area, perimeter, and edge information
These features convert images into numerical data suitable for ML algorithms.
4. Feature Selection
3
Relevant features are selected to reduce dimensionality and remove redundant information.
This improves classification accuracy and reduces computation time.
5. Dataset Splitting
The dataset is divided into:
Training set – used to train the model
Testing set – used to evaluate model performance
A standard split ratio (e.g., 80% training and 20% testing) is followed.
6. Model Training
Multiple machine learning algorithms are trained using the extracted features:
Support Vector Machine (SVM)
Random Forest
K-Nearest Neighbors (KNN)
Each model learns patterns that distinguish different skin lesion classes.
7. Model Evaluation
The trained models are evaluated using:
Accuracy
Precision
Recall
F1-score
Confusion Matrix
The performance of each algorithm is compared.
8. Model Selection
The best-performing model is selected based on evaluation metrics and overall reliability.
4
Expected Outcomes
The machine learning model demonstrates high accuracy in classifying skin cancer types on
the test dataset. Performance is evaluated using accuracy, confusion matrix, precision, recall,
and F1-score. The results indicate that machine learning techniques are effective for skin
cancer classification.
Applications
• Automated skin cancer screening systems
• Clinical decision support tools
• Early detection of skin diseases
• Telemedicine platforms
Conclusion
This project successfully demonstrates the application of machine learning techniques for
skin cancer classification. The system provides a reliable and efficient diagnostic aid that can
support dermatologists in early detection. Future improvements may include the use of
advanced feature extraction techniques and hybrid ML models.
Referencesss
1. Tschandl, P., Rosendahl, C., & Kittler, H. (2018).
Skin Cancer MNIST: HAM10000 – A large collection of dermatoscopic images of
common pigmented skin lesions.
Dataset available at Kaggle:
[Link]
2. International Skin Imaging Collaboration (ISIC).
Public archive of dermoscopic images for skin lesion analysis.
Available at: [Link]
5
3. Esteva, A., Kuprel, B., Novoa, R. A., et al. (2017).
Dermatologist-level classification of skin cancer with deep neural networks., Nature.
Available at: [Link]
4. World Health Organization (WHO).
Skin cancer information, causes, and prevention guidelines.
Available at: [Link]
radiation
5. Scikit-learn Developers.
Scikit-learn: Machine Learning in Python.
Available at: [Link]
6. Python Software Foundation.
Official Python documentation for data processing and machine learning
implementation.
Available at: [Link]
7. GeeksforGeeks.
Tutorials on machine learning algorithms and skin cancer detection concepts.
Available at: [Link]
learning/