Training.py

import os,cv2
import numpy as np
from PIL import Image

recognizer = cv2.createLBPHFaceRecognizer();
path = 'Data1'

def img(path):
    imgpaths = [os.path.join(path,f) for f in os.listdir(path)] #getting images path
    #print imgpaths
    
    faces=[]
    users=[]
    for imgpath in imgpaths:
        faceimg = Image.open(imgpath).convert('L') #converting to grayscale
        facenp = np.array(faceimg,'uint8')
        user=int(os.path.split(imgpath)[-1].split('.')[0]) #getting image ids
        faces.append(facenp)
        print user
        users.append(user)
        #cv2.imshow("Traning",facenp)
        cv2.waitKey(10)
    return users,faces


users , faces = img(path)
recognizer.train(faces,np.array(users)) #training
recognizer.save('TrainedDataFolder/TraningData.yml') #saving histogram data
cv2.destroyAllWindows()