app.py

import gradio as gr
import numpy as np
import tensorflow as tf

# Load your model
model = tf.keras.models.load_model("model.h5")  # Ensure this is the correct path to your model

def predict(image):
    # Preprocess the image for prediction
    image = tf.image.resize(image, [224, 224])  # Change to your model's expected size
    image = np.expand_dims(image, axis=0)  # Add batch dimension
    predictions = model.predict(image)  # Get model predictions
    
    # Assuming your model outputs probabilities for binary classification
    # The first output is the probability of class 0 (no tumor), 
    # and the second output is the probability of class 1 (tumor)
    no_tumor_confidence = predictions[0][0]  # Probability of no tumor
    tumor_confidence = predictions[0][1]      # Probability of tumor

    # Create a response with confidence scores
    if tumor_confidence > no_tumor_confidence:
        result = {
            "prediction": "Tumor Detected",
            "confidence": float(tumor_confidence)
        }
    else:
        result = {
            "prediction": "No Tumor Detected",
            "confidence": float(no_tumor_confidence)
        }

    return result

# Create a Gradio interface
iface = gr.Interface(fn=predict, inputs="image", outputs="json")

# Launch the Gradio interface
iface.launch()