AI-Powered Eye Disease Detection & Colour Vision Testing

OculusAI is an academic demonstration project showcasing the application of deep learning technology in medical imaging, specifically for detecting eye diseases from retinal fundus images and assessing colour vision deficiencies.

This project demonstrates how convolutional neural networks can be trained to assist in identifying common eye conditions and colour blindness patterns, providing a comprehensive proof-of-concept for AI in ophthalmology and vision science.

How It Works

Eye Disease Detection

1

Upload a retinal fundus image (JPEG, PNG, or GIF format)

2

Image is preprocessed and resized to 256×256 pixels

3

CNN model analyzes the image and predicts disease class

4

Get detailed results with confidence scores for all 4 classes

Colour Blindness Test

1

Take an interactive Ishihara-style colour vision test

2

AI model predicts the correct digit in each colour plate (99.5% accuracy)

3

System analyzes error patterns across 4 colour type variations

4

Receive probability-based diagnosis for Protan/Deutan deficiencies

Technology Stack

Dual AI Models

Disease detection + Ishihara digit recognition (99.5% accuracy)

Modern Frontend

Next.js 16 with React & TypeScript

Fast Backend

Flask API with multi-model inference

Advanced Analysis

Probability-based colour blindness diagnosis

Detection Capabilities

Eye Diseases (Retinal Analysis)

Cataract - Lens clouding

Diabetic Retinopathy - Retinal blood vessel damage

Glaucoma - Optic nerve damage

Normal - Healthy eye

Colour Vision Deficiencies (Ishihara Test)

Deuteranomaly - Reduced green sensitivity (mild)

Deuteranopia - Complete green colour blindness (moderate to strong)

Protanomaly - Reduced red sensitivity (mild)

Protanopia - Complete red colour blindness (moderate to strong)

Project Purpose

This is an academic project created to demonstrate the potential of AI in medical imaging. It serves as a proof-of-concept for how deep learning can be applied to assist healthcare professionals in early disease detection.

The project showcases modern web technologies, machine learning workflows, and user-centered design principles for medical applications.

Important Notice

This application is a demonstration project for educational and research purposes only. It is not intended for clinical use or as a substitute for professional medical diagnosis. Always consult qualified ophthalmologists for actual eye health concerns and medical advice.