buildface.py

import cv2,os,glob
import numpy as np
from time import sleep
def bface():

    def saveImg(image,index):
        filename="images/"+"/face{:03d}.jpg".format(index)
        cv2.imwrite(filename,image)
        
    index=1
    total = 100

    
    os.mkdir("images/")
    face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades+"haarcascade_frontalface_alt2.xml")
    cap = cv2.VideoCapture(0)
    cv2.namedWindow('video',cv2.WINDOW_NORMAL)
    while index>0:
        ret,frame = cap.read()
        frame = cv2.flip(frame,1)
        gray = cv2.cvtColor(frame,cv2.COLOR_BGR2GRAY)
        faces = face_cascade.detectMultiScale(gray,1.1,3)
        for(x,y,w,h) in faces:
                frame = cv2.rectangle(frame,(x,y),(x+w,y+h),(0,255,0),3)
                image = cv2.resize(gray[y:y + h,x: x+w],(400,400))
                saveImg(image,index)
                sleep(0.1)
                index+=1
                if index > total:
                    print("done!")

                    index = -1
                    break
                    
                    
        cv2.imshow('video',frame)
        cv2.waitKey(1)
    cap.release()
    cv2.destroyAllWindows()
    images=[]
    labels=[]
    labelstr=[]
    count=0
    dirs = os.listdir('images')
    for d in dirs:
        if os.path.isdir('images/'):
            files = glob.glob("images/"+"/*.jpg")
            for filename in files:
                img = cv2.imread(filename,cv2.COLOR_BGR2GRAY)
                images.append(img)
                labels.append(count)
            labelstr.append(d)
            count += 1


    print("start create...")
    model = cv2.face.LBPHFaceRecognizer_create()
    model.train(np.asarray(images),np.asarray(labels))
    model.save('faces_LBPH.yml')
    print("create done!")