app.py

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

# Load your trained model
model = load_model("model.h5")

# Define your class labels
class_labels = ['Angry', 'Disgust', 'Fear', 'Happy', 'Sad', 'Surprise', 'Neutral']

# Function to preprocess and predict the emotion
def predict_expression(image):
    if image is None:
        return "No image provided."

    # Convert to grayscale
    gray_image = image.convert("L")

    # Resize to match model input
    resized_image = gray_image.resize((48, 48))

    # Normalize and reshape
    img_array = np.array(resized_image) / 255.0
    img_array = img_array.reshape(1, 48, 48, 1)

    # Predict
    prediction = model.predict(img_array)
    predicted_label = class_labels[np.argmax(prediction)]

    return predicted_label

# Toggle function to switch between upload and camera
def toggle_input(method):
    return (
        gr.update(visible=(method == "Upload")),
        gr.update(visible=(method == "Camera"))
    )

# Gradio interface
with gr.Blocks() as app:
    gr.Markdown("## 😃 Facial Expression Recognition")

    input_method = gr.Radio(["Upload", "Camera"], label="Select Input Method", value="Upload")

    upload_image = gr.Image(label="Upload Image", type="pil", visible=True)
    camera_image = gr.Image(label="Camera Input", type="pil", visible=False, sources=["webcam"])

    predict_button = gr.Button("Predict Expression")

    output_label = gr.Label(label="Predicted Expression")

    input_method.change(fn=toggle_input, inputs=input_method, outputs=[upload_image, camera_image])

    predict_button.click(
        fn=predict_expression,
        inputs=[upload_image, camera_image],
        outputs=output_label,
        preprocess=False  # We'll handle the logic manually
    )

app.launch()