AI PROJECT CLASS X/XII

Project 4:

AI-Based Emotion Recognition System

Objective: Build a system that can recognize emotions from facial expressions using AI.

Problem Statement: Understanding human emotions is crucial in many fields, including mental health care, customer service, and human-computer interaction. However, detecting emotions in real-time through traditional means, such as verbal communication or self-reporting, is often inefficient and subjective. This project aims to create an AI-powered emotion recognition system that uses facial expression analysis to automatically detect and classify emotions from real-time video inputs. By leveraging deep learning and computer vision, the system will enhance emotional awareness in applications like customer support systems, virtual learning environments, and therapeutic monitoring, providing a more responsive and empathetic interaction experience.

Key Tools:

• Use facial recognition libraries like face_recognition or dlib for detecting facial features.
• Use a pre-trained model or train a CNN (Convolutional Neural Network) on an emotion dataset like FER2013.
• Utilize Python libraries like OpenCV for video capture and real-time emotion detection.

Features:

• Detect emotions like happiness, sadness, anger, etc., in real-time from live webcam feeds.
• Display emotion results on the screen.
• Store the detected emotions along with a timestamp in a log file for further analysis.

Python Code:

import cv2
import numpy as np
from keras.models import load_model
from keras.preprocessing.image import img_to_array

# Load pre-trained emotion recognition model (FER2013)
model = load_model('emotion_model.hdf5')  # Use a pre-trained model like FER2013

# Emotion classes (can vary depending on the model)
emotion_classes = ['Angry', 'Disgust', 'Fear', 'Happy', 'Sad', 'Surprise', 'Neutral']

# Initialize webcam
cap = cv2.VideoCapture(0)

# Load face detector (Haar cascade)
face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')

while True:
    ret, frame = cap.read()
    if not ret:
        break

    # Convert frame to grayscale (required for face detection)
    gray_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

    # Detect faces
    faces = face_cascade.detectMultiScale(gray_frame, scaleFactor=1.3, minNeighbors=5)

    for (x, y, w, h) in faces:
        # Extract face ROI
        face_roi = gray_frame[y:y+h, x:x+w]
        face_roi = cv2.resize(face_roi, (48, 48))  # Model expects 48x48 input
        face_roi = face_roi.astype('float') / 255.0
        face_roi = img_to_array(face_roi)
        face_roi = np.expand_dims(face_roi, axis=0)

        # Predict emotion
        prediction = model.predict(face_roi)[0]
        max_index = np.argmax(prediction)
        emotion = emotion_classes[max_index]

        # Draw bounding box and emotion label
        cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 0), 2)
        cv2.putText(frame, emotion, (x, y-10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (0, 255, 0), 2)

    # Show the frame
    cv2.imshow('Emotion Recognition', frame)

    # Press 'q' to exit
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()

Key Points:

• This code detects facial emotions using a pre-trained model like FER2013, which you can train or download.
• The emotion classes list may change depending on the model used. Modify this to match your model’s output.