import streamlit as st
import cv2
import numpy as np
from tensorflow.keras.models import load_model


model = load_model("FER-CNN.h5")

# Define emotion labels
emotion_labels = ['Angry', 'Fear', 'Happy', 'Neutral', 'Sad', 'Surprise']

def detect_emotion(frame):
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
    faces = face_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5, minSize=(30, 30))
    
    for (x, y, w, h) in faces:
        roi_gray = gray[y:y + h, x:x + w]
        roi_gray = cv2.resize(roi_gray, (48, 48), interpolation=cv2.INTER_AREA)
        roi_gray = roi_gray.astype('float') / 255.0
        roi_gray = np.expand_dims(roi_gray, axis=-1)      
        roi_gray = np.expand_dims(roi_gray, axis=0)  

        prediction = model.predict(roi_gray)[0]
        emotion = emotion_labels[prediction.argmax()]

        # Draw the rectangle and label on the frame
        cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 255), 2)
        cv2.putText(frame, emotion, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 2)
    
    return frame

st.title("Real-Time Face Emotion Recognition")

run = st.checkbox("Start Webcam")

FRAME_WINDOW = st.image([])

camera = cv2.VideoCapture(0)  # Start the webcam

if run:
    while True:
        ret, frame = camera.read()
        if not ret:
            st.error("Failed to capture video.")
            break

        frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
        frame = detect_emotion(frame)
        FRAME_WINDOW.image(frame, channels="RGB")
else:
    camera.release()
    cv2.destroyAllWindows()