# app.py import streamlit as st import cv2 import numpy as np import tensorflow as tf import time import os # --- Streamlit Page Configuration (MUST BE THE FIRST STREAMLIT COMMAND) --- st.set_page_config(page_title="Real-time Emotion Recognition", layout="wide") # --- 1. Load Model and Face Detector (Cached for Performance) --- @st.cache_resource def load_emotion_model(): model_path = 'models/emotion_model_best.h5' # Path to your trained model if not os.path.exists(model_path): st.error(f"Error: Model file not found at {model_path}. Please ensure training was successful and the file exists.") st.stop() try: model = tf.keras.models.load_model(model_path) return model except Exception as e: st.error(f"Error loading model from {model_path}: {e}") st.stop() @st.cache_resource def load_face_detector(): cascade_path = 'haarcascade_frontalface_default.xml' # Path to your Haar Cascade file if not os.path.exists(cascade_path): st.error(f"Error: Haar Cascade file not found at {cascade_path}.") st.markdown("Please download `haarcascade_frontalface_default.xml` from:") st.markdown("[https://github.com/opencv/opencv/blob/4.x/data/haarcascades/haarcascade_frontalface_default.xml](https://github.com/opencv/opencv/blob/4.x/data/haarcascades/haarcascade_frontalface_default.xml)") st.markdown("And place it in a `cascades` folder next to `app.py`.") st.stop() face_cascade = cv2.CascadeClassifier(cascade_path) if face_cascade.empty(): st.error(f"Error: Could not load Haar Cascade classifier from {cascade_path}. Check file integrity.") st.stop() return face_cascade # Load the model and face detector when the app starts model = load_emotion_model() face_detector = load_face_detector() # --- 2. Define Constants and Labels --- IMG_HEIGHT = 48 IMG_WIDTH = 48 emotion_labels = ['angry', 'disgust', 'fear', 'happy', 'neutral', 'sad', 'surprise'] label_colors = { 'angry': (0, 0, 255), # BGR Red 'disgust': (0, 165, 255), # BGR Orange 'fear': (0, 255, 255), # BGR Yellow 'happy': (0, 255, 0), # BGR Green 'neutral': (255, 255, 0), # BGR Cyan 'sad': (255, 0, 0), # BGR Blue 'surprise': (255, 0, 255) # BGR Magenta } # --- 3. Streamlit App Layout --- st.title("Live Facial Emotion Recognition") st.markdown(""" This application uses a deep learning model (trained on FER-2013) to detect emotions from faces in real-time. It requires access to your computer's webcam. """) stframe = st.empty() st_status = st.empty() col1, col2 = st.columns([1,1]) with col1: start_button = st.button("Start Camera", key="start_camera") with col2: stop_button = st.button("Stop Camera", key="stop_camera") # Initialize session state for camera control and performance tracking if "camera_started" not in st.session_state: st.session_state.camera_started = False if "cap" not in st.session_state: st.session_state.cap = None if "last_process_time" not in st.session_state: st.session_state.last_process_time = 0.0 # --- Performance Configuration --- DESIRED_FPS = 15 # Aim for 15 frames per second for processing FRAME_INTERVAL_SECONDS = 1.0 / DESIRED_FPS FACE_DETECTION_DOWNSCALE = 0.5 # Scale factor for face detection (e.g., 0.5 means half size) # --- 4. Main Camera Loop Logic --- if start_button: st.session_state.camera_started = True if stop_button: st.session_state.camera_started = False st_status.info("Camera stopped.") if st.session_state.cap is not None and st.session_state.cap.isOpened(): st.session_state.cap.release() st.session_session.cap = None stframe.empty() # Updated: use_container_width instead of use_column_width stframe.image(np.zeros((480, 640, 3), dtype=np.uint8), channels="RGB", use_container_width=True) if st.session_state.camera_started: st_status.info("Starting camera... Please allow camera access if prompted.") if st.session_state.cap is None or not st.session_state.cap.isOpened(): st.session_state.cap = cv2.VideoCapture(0, cv2.CAP_DSHOW) if not st.session_state.cap.isOpened(): st_status.error("Failed to open camera. Please check if it's connected and not in use.") st.session_state.camera_started = False st.stop() while st.session_state.camera_started: ret, frame = st.session_state.cap.read() if not ret: st_status.error("Failed to read frame from camera. It might be disconnected or an error occurred.") st.session_state.camera_started = False break current_time = time.time() if current_time - st.session_state.last_process_time >= FRAME_INTERVAL_SECONDS: st.session_state.last_process_time = current_time gray_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) small_frame = cv2.resize(gray_frame, (0, 0), fx=FACE_DETECTION_DOWNSCALE, fy=FACE_DETECTION_DOWNSCALE) faces = face_detector.detectMultiScale(small_frame, scaleFactor=1.1, minNeighbors=5, minSize=(30, 30)) original_faces = [] for (x, y, w, h) in faces: x_orig = int(x / FACE_DETECTION_DOWNSCALE) y_orig = int(y / FACE_DETECTION_DOWNSCALE) w_orig = int(w / FACE_DETECTION_DOWNSCALE) h_orig = int(h / FACE_DETECTION_DOWNSCALE) original_faces.append((x_orig, y_orig, w_orig, h_orig)) for (x, y, w, h) in original_faces: cv2.rectangle(frame, (x, y), (x+w, y+h), (255, 0, 0), 2) face_roi = gray_frame[max(0, y):min(gray_frame.shape[0], y+h), max(0, x):min(gray_frame.shape[1], x+w)] if face_roi.size == 0: continue face_roi = cv2.resize(face_roi, (IMG_WIDTH, IMG_HEIGHT)) face_roi = np.expand_dims(face_roi, axis=0) face_roi = np.expand_dims(face_roi, axis=-1) face_roi = face_roi / 255.0 predictions = model.predict(face_roi, verbose=0)[0] emotion_index = np.argmax(predictions) predicted_emotion = emotion_labels[emotion_index] confidence = predictions[emotion_index] * 100 text_color = label_colors.get(predicted_emotion, (255, 255, 255)) text = f"{predicted_emotion} ({confidence:.2f}%)" text_y = y - 10 if y - 10 > 10 else y + h + 20 cv2.putText(frame, text, (x, text_y), cv2.FONT_HERSHEY_SIMPLEX, 0.9, text_color, 2, cv2.LINE_AA) frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB) # Updated: use_container_width instead of use_column_width stframe.image(frame_rgb, channels="RGB", use_container_width=True) time.sleep(0.001) # Small sleep to yield control, can be adjusted or removed if st.session_state.cap is not None and st.session_state.cap.isOpened(): st.session_state.cap.release() st.session_state.cap = None st_status.info("Camera released.")