day-8_face_recogn..py

import joblib
import cv2 
import numpy as np
import matplotlib.pyplot as plt

cam = cv2.VideoCapture(1)
path = "E:\\Coding\\AIML\\haarcascade_frontalface_default.xml"
# face detector 
face_detector = cv2.CascadeClassifier(path)

model_path = 'E:\\Coding\\AIML\\data_files\\orl_face_model_2.pkl'
face_model = joblib.load(model_path)

frame=True
count = 0
while(frame):
    ret,im = cam.read()
    im_new = cv2.resize(im, (512,512))
    # covert the color (BGR) into grayscale
    gray_im = cv2.cvtColor(im_new,cv2.COLOR_BGR2GRAY)
    # run your classifier on the image
    faces = face_detector.detectMultiScale(gray_im,scaleFactor=1.1,minNeighbors=10)
    # disply the bounding box on all the faces
    for (dx,dy,w,h) in faces:
        cv2.rectangle(im_new, (dx,dy),(dx+w,dy+h),(0,0,255),2)
        cropped_im = cv2.resize((gray_im[dy-20:(dy+h)+40,dx:(dx+w)]),(92,112))
        lb = face_model.predict((cropped_im.reshape(1,-1)))
        cv2.putText(im_new,'user: '+str(lb[0]),(dx-5,dy-5),cv2.FONT_HERSHEY_SIMPLEX,1,color=(255,0,0),thickness=2)


    cv2.imshow('camera live feed', im_new)
    # desired button of your choice 
    if cv2.waitKey(1) & 0xFF == ord('q'):
        frame=False
        break
  

cam.release()
cv2.destroyAllWindows()