app.py

import gradio as gr
import tensorflow as tf
import numpy as np
from tensorflow.keras.preprocessing import image
from PIL import Image

# Load the trained model
model = tf.keras.models.load_model("my_keras_model.h5")

# Define image size based on the model's input requirement
image_size = (224, 224)

# Function to make predictions
def predict_image(img):
    img = img.resize(image_size)  # Resize image to model's expected size
    img_array = image.img_to_array(img)
    img_array = np.expand_dims(img_array, axis=0) / 255.0  # Normalize
    prediction = model.predict(img_array)

    # Assuming binary classification (fractured or normal)
    class_names = ['Fractured', 'Normal']
    predicted_class = class_names[int(prediction[0] > 0.5)]  # Threshold at 0.5

    return f"Prediction: {predicted_class} (Confidence: {prediction[0][0]:.2f})"

# Preloaded images for testing
sample_images = [
    ("fracture1.jpg", "Fractured Example"),
    ("fracture2.jpg", "Fractured Example"),
    ("normal1.jpg", "Normal Example"),
    ("normal2.jpg", "Normal Example"),
]

# Define Gradio Interface
interface = gr.Interface(
    fn=predict_image,
    inputs=gr.Image(type="pil"),
    outputs=gr.Textbox(),
    examples=sample_images,  # Preloaded images for testing
    title="Bone Fracture Detection",
    description="""
    <h1 style='color: blue;'>Bone Fracture Detection Model</h1>
    <p>This AI model predicts whether a given X-ray image shows a fracture or not.</p>
    <p><b>Upload an image</b> or select from the provided samples to get a prediction.</p>
    """,
)

# Launch the Gradio app
if __name__ == "__main__":
    interface.launch()