Model Card for TransferLearningDR: Diabetic Retinopathy Detector

TransferLearningDR is a convolutional neural network (CNN) model built on EfficientNetB3 and trained to classify retinal fundus images into two categories: No DR and Diabetic Retinopathy (DR). The model leverages transfer learning from ImageNet weights to accelerate training and improve generalization.

⚠️ Disclaimer: This model is intended for research and educational purposes only. It is not a substitute for professional medical diagnosis. Always consult a licensed healthcare provider for medical decisions.

---

Model Details

Key Features:

• Binary classification of retinal images (No DR vs DR)
• Transfer learning using EfficientNetB3 pretrained on ImageNet
• Input size: 224x224 RGB images, normalized
• Custom dense head with batch normalization, dropout, and L1/L2 regularization
• Evaluated on separate validation and test sets with high accuracy

Skills & Technologies Used:

• TensorFlow / Keras for model design & training
• ImageDataGenerator for preprocessing and optional augmentation
• EfficientNetB3 for feature extraction
• Matplotlib & Seaborn for visualization
• Hugging Face Hub for model hosting
• Kaggle for dataset and training environment

---

• Developed by: Rawan Alwadeya
• Model type: Convolutional Neural Network (CNN) with Transfer Learning
• Language(s): N/A (Image model)
• License: MIT

---

Uses

This model can be used for:

• Research in diabetic retinopathy detection and AI-assisted diagnosis
• Educational demonstrations of transfer learning on medical images
• Portfolio projects showcasing preprocessing, model training, and deployment

---

👩‍💻 Author

Rawan Alwadeya
AI Engineer | Generative AI Engineer | Data Scientist

• 📧 Email: [email protected]
• 🌐 LinkedIn Profile

---

Example Usage

import numpy as np
from keras.models import load_model
from PIL import Image
import requests
from io import BytesIO
from huggingface_hub import hf_hub_download

# Load the model from Hugging Face Hub
model_path = hf_hub_download(repo_id="RawanAlwadeya/TransferLearningDR", filename="TransferLearningDR.h5")
model = load_model(model_path)

# Preprocess an example image
def preprocess_image(image):
    IMG_SIZE = (224, 224)
    image = image.convert("RGB")  
    image = image.resize(IMG_SIZE)
    img_array = np.array(image) / 255.0
    img_array = np.expand_dims(img_array, axis=0)  # add batch dimension
    return img_array

# Example: load an image from URL
url = "https://example.com/sample_fundus.jpg"
response = requests.get(url)
image = Image.open(BytesIO(response.content))

img_array = preprocess_image(image)
prediction = model.predict(img_array)[0][0]

if prediction >= 0.5:
    print("⚠️ Diabetic Retinopathy likely detected")
else:
    print("✅ Likely No DR")