Final Year Project (FYP)

About SmartScan+

SmartScan+ is an AI-based medical screening system developed as a Final Year Project (FYP) for the BS Computer Science program. The project focuses on early risk detection of skin lesions and anemia using computer vision and deep learning techniques.

Project Overview

The project demonstrates the practical application of artificial intelligence in healthcare by integrating machine learning models with mobile and web-based systems to support preliminary medical screening.

AI-Based Screening

Deep learning models analyze medical images to estimate the likelihood of skin lesion categories and anemia-related visual indicators.

Research-Oriented Datasets

The models are trained using publicly available and research-backed medical datasets to ensure reproducibility and academic validity.

End-to-End System

The solution includes mobile image acquisition, backend processing, AI inference, and a web-based results dashboard.

Technical Focus Areas

  • Computer vision for medical image analysis
  • Deep learning model training and evaluation
  • Mobile application development
  • Web-based visualization dashboard
  • Secure database and backend integration

System Scope & Objectives

SmartScan+ is designed as a decision-support and preliminary screening tool. It aims to assist users and healthcare professionals by providing early risk indications rather than definitive medical diagnoses.

Core Objectives

  • Early screening of common skin lesion categories
  • Non-invasive anemia risk estimation
  • Improved accessibility to preliminary health screening
  • Demonstration of AI integration in healthcare systems

Intended Users

  1. General users seeking early health awareness
  2. Healthcare professionals for decision support

System Architecture

The system follows a modular and scalable architecture separating data capture, AI processing, backend services, and user interfaces.

Capture
Mobile photo capture (dermoscopy / eye)
Preprocess
Image normalization & quality checks
Inference
On-device or server-side model evaluation
Backend
Auth, API, logging, and secure storage
Storage
Encrypted results and research datasets

Limitations & Ethical Considerations

SmartScan+ provides probabilistic outputs based on trained datasets. The system is not a substitute for professional medical diagnosis and should be used responsibly.

  • Model accuracy depends on dataset diversity and image quality
  • Performance may vary across lighting conditions and skin tones
  • User data privacy and security are critical design priorities
  • Clinical validation is required for real-world deployment