PopularMajor Project
Artificial Intelligence
Face Recognition Attendance System
Camera-based attendance with liveness detection and analytics
- Python
- OpenCV
- FaceNet
- Flask
- +2
₹4,900₹5,999
4.8/301 reviews
Artificial Intelligence · Major Project
Diabetic Retinopathy Grading with Deep Learning
Grades retinopathy severity from fundus photographs on a 0-4 scale
3-4 weeks build
Overview
Fine-tunes a pretrained CNN on the APTOS 2019 fundus dataset to grade retinopathy from no disease to proliferative. Images go through circular cropping and contrast normalisation first, the model is scored with quadratic weighted kappa, and a web page shows the grade alongside a Grad-CAM heatmap.
What makes it stand up
- 01APTOS 2019 blindness-detection images with five severity grades
- 02Fundus preprocessing: circle crop, Ben Graham colour normalisation
- 03Transfer learning with class-weighted loss for rare severe grades
- 04Quadratic weighted kappa and confusion matrix in the report
- 05Academic screening prototype, not a replacement for an ophthalmologist
Module breakdown
- 01
Preprocessing
Crop, resize and normalise fundus images consistently.
- 02
Augmentation
Rotation, flips and brightness jitter suited to retinal images.
- 03
Model
Pretrained EfficientNet or ResNet backbone with a grading head.
- 04
Explainability
Grad-CAM overlays highlighting lesion regions.
- 05
Web app
Upload page returning grade, confidence and heatmap.
After this, you will be able to
- Explain why quadratic weighted kappa suits ordinal grades
- Defend your handling of class imbalance across severity levels
- Demo grading and heatmaps on unseen fundus images
Technology stack
- Python
- PyTorch
- OpenCV
- FastAPI
- React
- Level
- Major Project
- Domain
- Artificial Intelligence
- Build time
- 3-4 weeks
- Difficulty
You receive
- Complete, commented source code
- 60-90 page project report (IEEE format)
- Editable presentation deck
- Architecture, ER and UML diagrams
- Local setup and deployment guide
- Sample dataset or seed data
- Viva question bank with answers
- One 45-minute walkthrough call
- 30 days of doubt-clearing support