main.py

import cv2
from tensorflow import keras
import numpy as np
import h5py

model = keras.models.load_model('trial_model.h5')
print('\n\n')
print(model.summary())
########################################
face_cascade = cv2.CascadeClassifier("/home/tejal/Documents/soc_mini_project/haarcascade.xml")

cap = cv2.VideoCapture(0)
outputToEmotionMap = {
    0: 'angry',
    1: 'disgusted',
    2: 'in fear',
    3: 'happy',
    4: 'neutral',
    5: 'sad',
    6: 'surprised'
}

while cap.isOpened():
    __, img = cap.read()

    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

    faces = face_cascade.detectMultiScale(gray, scaleFactor = 1.3, minNeighbors = 3)
    x1 = 0
    y1 = 0
    w1 = 0
    h1 = 0

    for (x,y,w,h) in faces:
        cv2.rectangle(img, (x, y), (x+w, y+h), (255,0,0), 3)
        x1 = x
        y1 = y
        w1 = w
        h1 = h

    cv2.imshow("img", img)
    ######################################
    inputImage = np.reshape(cv2.resize(gray[y1:y1 + h1, x1:x1 + w1], (48,48)),(-1,48,48,1))
    print("SHAPE IS", np.shape(inputImage))

    # predicting with help of model - taking one with max probability
    predictions = model.predict(inputImage)
    maxProbabilityPredict = np.argmax(predictions)

    # adjusting message
    message = 'You are {}'.format(
        outputToEmotionMap[maxProbabilityPredict]
    )
    print("you are", message)

    ##################
    font = cv2.FONT_HERSHEY_SIMPLEX

    # org
    org = (x,y)

    # fontScale
    fontScale = 1

    # Blue color in BGR
    color = (0, 255, 0)

    # Line thickness of 2 px
    thickness = 2

    # Using cv2.putText() method
    image = cv2.putText(img, message, org, font,
                        fontScale, color, thickness, cv2.LINE_AA)
    cv2.imshow("emo_img", img)
    if cv2.waitKey(1) == ord("q"):
        break

cap.release()