recognizer.py 文件源码

 复制代码def __init__(self):

        face_cascade = cv2.CascadeClassifier('Haar/haarcascade_frontalcatface.xml')
        eye_cascade = cv2.CascadeClassifier('Haar/haarcascade_eye.xml')

        recognise = cv2.face.createEigenFaceRecognizer(15, 4000)  # creating EIGEN FACE RECOGNISER
        recognise.load("Recogniser/trainingDataEigan.xml")  # Load the training data

        # -------------------------     START THE VIDEO FEED ------------------------------------------
        cap = cv2.VideoCapture(0)  # Camera object
        # cap = cv2.VideoCapture('TestVid.wmv')   # Video object
        ID = 0
        while True:
            ret, img = cap.read()  # Read the camera object
            gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)  # Convert the Camera to gray
            faces = face_cascade.detectMultiScale(gray, 1.3, 5)  # Detect the faces and store the positions
            for (x, y, w, h) in faces:  # Frames  LOCATION X, Y  WIDTH, HEIGHT
                # ------------ BY CONFIRMING THE EYES ARE INSIDE THE FACE BETTER FACE RECOGNITION IS GAINED ------------------
                gray_face = cv2.resize((gray[y: y + h, x: x + w]), (110, 110))  # The Face is isolated and cropped
                eyes = eye_cascade.detectMultiScale(gray_face)
                for (ex, ey, ew, eh) in eyes:
                    ID, conf = recognise.predict(gray_face)  # Determine the ID of the photo
                    NAME = NameFind.ID2Name(ID, conf)
                    NameFind.DispID(x, y, w, h, NAME, gray)
            cv2.imshow('EigenFace Face Recognition System', gray)  # Show the video
            if cv2.waitKey(1) & 0xFF == ord('q'):  # Quit if the key is Q
                break
        cap.release()
        cv2.destroyAllWindows()