File size: 943 Bytes
79d52c5
c5e4951
5305212
 
 
2d57b8b
9730dc3
2d57b8b
79d52c5
 
 
 
c5e4951
5305212
c5e4951
 
 
 
 
 
 
 
5305212
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
---
title: AI-Powered Diabetic Retinopathy Grading System from Retinal Fundus Images Using Deep Learning
emoji: 👁️
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 5.29.0
app_file: app.py
pinned: false
license: mit
---

# Diabetic Retinopathy Grading AI (v9)

An AI-powered clinical diagnostic tool for grading Diabetic Retinopathy from retinal fundus photographs. 
Powered by our latest **v9** training phase.

## Performance Metrics (v9)
- **Validation QWK**: 0.9167
- **Test QWK**: 0.8954
- **Binary Referable-DR Accuracy**: 91.92%
- **Test AUROC (macro)**: 0.8885

## Features
- **5-Grade ICDR Classification** using EfficientNetV2-M + CORAL ordinal regression
- **Grad-CAM++ Visual Explanations** highlighting lesion regions
- **Fundus Gating** — rejects non-retinal images automatically
- **4-View TTA** (Test-Time Augmentation) for improved robustness
- **Premium Material Design 3 UI** with real-time processing metrics