You can deploy this model and start earning money today!

seyidyildiz/
retina_disease_risk

Name: retina_disease_risk
Availability: InStock

apache-2.0

Views

No views yet

Retinal Disease Risk Detection

Model Description

This deep learning model automatically detects the risk of retinal disease from fundus images. It classifies a patient as either "No Disease Risk" or "Disease Risk Present".
The goal is to assist doctors with early diagnosis as a preliminary screening tool.

Intended Use

Diagnostic support in a clinical setting.
Helps prioritize cases and streamline evaluation of at-risk patients.
Not a standalone medical diagnostic tool.

Model Architecture and Training

Model Name: Retinal Disease Risk Detection Model
Architecture: Convolutional Neural Network (CNN)
Training Dataset: RFMiD (Retinal Fundus Image Multidisease)
Data Preparation: Images resized to 224x224 and normalized (0-1)
Optimizer: Adam, learning rate = 0.0001
Techniques: Data augmentation, class weights, early stopping to prevent overfitting

Model Performance

Class	Precision	Recall	F1-Score	Support
No Risk	0.63	0.61	0.62	134
Risk Present	0.90	0.90	0.90	506
Weighted Avg	0.84	0.84	0.84	640

Overall Accuracy: 84.22%
The model performs well on "Disease Risk Present" but has lower recall for "No Risk".

Limitations and Ethical Considerations

Not a diagnostic tool. Final diagnosis must be by a qualified healthcare professional.

Lower recall for "No Risk" could misclassify healthy individuals.

Model accuracy depends on the quality and diversity of the training data; performance may vary across demographics and imaging conditions.

Contact

Name: Seyid Yıldız

Email: seyidyildiz416@gmail.com

LinkedIn: https://www.linkedin.com/in/seyid-yıldız-310091349

Installation

bash

pip install tensorflow opencv-python numpy
import numpy as np
import cv2
from tensorflow.keras.models import load_model

# Load the model
model = load_model("retina_disease_risk.h5")

# Load and preprocess an image
img_path = 'new_image.png'
img = cv2.imread(img_path)
img = cv2.resize(img, (224, 224))
img = np.expand_dims(img, axis=0) / 255.0

# Prediction
prediction = model.predict(img)
if prediction[0][0] > 0.5:
    print("Disease Risk Present")
else:
    print("No Disease Risk")

seyidyildiz/retina_disease_risk