app.py

import gradio as gr
import numpy as np
from tensorflow.keras.models import load_model
from PIL import Image

model = load_model("final/brain_tumor_fine_tune.h5")
model.load_weights("final/brain_tumor_fine_tune.h5")


class_names = [
    'glioma',
    'meningioma',
    'notumor',
    'pituitary',
]

def classify_tumor(input_image):
    
    img = Image.fromarray(input_image)
    resized_img = img.resize((299, 299))
    img = np.asarray(resized_img)
    img = np.expand_dims(img, axis=0)
    img = img / 255
    prediction = model.predict(img)
    
    confidences = {class_names[i]: float(prediction[0][i]) for i in range(len(class_names))}
    
    return {class_name: conf for class_name, conf in sorted(confidences.items(), key=lambda x: x[1], reverse=True)}

iface = gr.Interface(
    fn=classify_tumor,
    inputs=gr.Image(type="numpy"),
    outputs=gr.Label(num_top_classes=5),
    title="Brain Tumor Classification",
    description="Upload an image of a Brain to classify its tumor.",
    theme=gr.themes.Soft()
)

# Launch the interface
iface.launch(share=True)