adelekedaniel21/eye_disease_detection

1import gradio as gr2from tensorflow.keras.models import load_model3from skimage.transform import resize4import numpy as np5 6 7# Load the trained model8model = load_model("eye_model.h5")9 10# Define a function to process the image and make predictions11def classify_image(image):12    # Resize image to model's input shape13    resized_image = resize(image, (224, 224, 3))  # Resize to (224, 224, 3)14    resized_image = np.expand_dims(resized_image, axis=0)  # Add batch dimension15    predictions = model.predict(resized_image)  # Make prediction16    class_probabilities = predictions[0]  # Probabilities for each class17        # Get the top prediction indices and map them to class names18    list_index = np.argsort(class_probabilities)[::-1]  # Sort indices by probability (descending)19    data_dir_list = ['cataract', 'diabetes', 'glaucoma', 'normal', 'others']20    top_predictions = {21        data_dir_list[i]: round(class_probabilities[i], 2)22        for i in list_index[:5]23    }24 25    # Return top predictions26    return top_predictions27 28# Set up the Gradio interface29interface = gr.Interface(30    fn=classify_image,  # Function to run for predictions31    inputs=gr.Image(type="numpy"),  # Accept image input and return as NumPy array32    outputs=gr.Label(num_top_classes=3),  # Display top 3 predictions33    title="Eye Condition Classification",34    description="Upload an eye image to classify its condition."35)36 37# Launch the interface38if __name__ == "__main__":39    interface.launch()40