abhay0283/major-project repository overview

⁠🩺 Skin Cancer Detection System (BCA Major Project)

⁠📌 Project Description

The Skin Cancer Detection System is a deep learning-based web application developed as a BCA 6th Semester Major Project. The primary objective of this project is to assist in the early detection and classification of skin cancer by analyzing dermoscopic images using multiple Convolutional Neural Network (CNN) models.

This system integrates three different model architectures—a pretrained model, a custom-built CNN, and a hybrid model—to improve prediction robustness and enable comparative analysis. By leveraging these models, the application provides users with accurate predictions along with confidence scores, helping in understanding model performance.

The application is built using Flask for backend integration and provides a simple, interactive interface where users can upload skin lesion images and receive instant predictions.

⁠🎯 Objectives

• To develop an AI-based system for skin lesion classification

• To compare performance between:

  • Pretrained model (EfficientNetB3)
  • Custom CNN model
  • Hybrid deep learning model

• To create a user-friendly web interface for real-time predictions

• To demonstrate the application of deep learning in healthcare

⁠⚙️ Working Principle

1. User uploads a skin lesion image

2. Image undergoes preprocessing (resize, normalization)

3. Image is passed through:

   • Pretrained model (EfficientNetB3)
   • Custom CNN model
   • Hybrid model

4. Each model predicts the class label (e.g., Melanoma, Nevus, BKL)

5. Results are displayed with confidence scores and comparison

⁠🧠 Key Highlights

• Multi-model prediction system
• Hybrid model implementation (combining features of pretrained + custom)
• Real-time inference using Flask
• Visualization of training performance (accuracy, loss, etc.)
• Dockerized deployment for easy execution

⁠🛠️ Technologies Used

• Programming Language: Python
• Framework: Flask
• Deep Learning: TensorFlow / Keras
• Libraries: NumPy, Pillow
• Frontend: HTML, CSS, JavaScript
• Deployment: Docker

⁠📊 Models Used

⁠🚀 Deployment

The project is containerized using Docker and available on Docker Hub:

docker pull abhay0283/major-project:latest
docker run -d -p 5000:5000 abhay0283/major-project:latest

⁠📈 Applications

• Early detection support for skin cancer
• Educational tool for understanding AI in healthcare
• Research and experimentation with deep learning models

⁠⚠️ Disclaimer

This system is developed for academic and research purposes only and should not be used as a substitute for professional medical diagnosis.