如何让dlib库读取图片文件夹？

Question

我正在使用 python dlib 库作为计算面部特征成对偏差的代码。当时读取一张图片时它工作得很好，但问题是我有 1200 张这样的图片。

如何让我的代码使用 .jpg 和 .jpeg 文件的文件夹并输出一个值数组？

这是我的代码：

predictor = dlib.shape_predictor('/Users/ik/PycharmProjects/FaceSymmetry/venv'
                                 '/shape_predictor_68_face_landmarks.dat')
detector = dlib.get_frontal_face_detector()
image = cv2.imread("/Users/ik/Desktop/Papers/Thesis/Photos/britney.jpg")

faces = detector(image, 1)
for face in faces:
    landmarks = predictor(image, face)

    left_jaw_x = landmarks.part(4).x
    left_jaw_y = landmarks.part(4).y
    right_jaw_x = landmarks.part(12).x
    right_jaw_y = landmarks.part(12).y
    mid_jaw_x = landmarks.part(66).x
    mid_jaw_y = landmarks.part(66).y
    left_jaw_dist = math.sqrt((left_jaw_x - mid_jaw_x) ** 2 + (left_jaw_y - mid_jaw_y) ** 2)
    right_jaw_dist = math.sqrt((right_jaw_x - mid_jaw_x) ** 2 + (right_jaw_y - mid_jaw_y) ** 2)
    # Jaw-midpoint deviation, pairwise (right side from left side):
    jaw_dev = math.sqrt((left_jaw_dist - right_jaw_dist) ** 2)

    left_lip_x = landmarks.part(48).x
    left_lip_y = landmarks.part(48).y
    right_lip_x = landmarks.part(54).x
    right_lip_y = landmarks.part(54).y
    mid_lip_x = landmarks.part(66).x
    mid_lip_y = landmarks.part(66).y
    left_lip_dist = math.sqrt((left_lip_x - mid_lip_x) ** 2 + (left_lip_y - mid_lip_y) ** 2)
    right_lip_dist = math.sqrt((right_lip_x - mid_lip_x) ** 2 + (right_lip_y - mid_lip_y) ** 2)
    # Lip-midpoint deviation, pairwise:
    lip_dev = math.sqrt((left_lip_dist - right_lip_dist) ** 2)

    left_nose_x = landmarks.part(31).x
    left_nose_y = landmarks.part(31).y
    right_nose_x = landmarks.part(35).x
    right_nose_y = landmarks.part(35).y
    mid_nose_x = landmarks.part(33).x
    mid_nose_y = landmarks.part(33).y
    left_nose_dist = math.sqrt((left_nose_x - mid_nose_x) ** 2 + (left_nose_y - mid_nose_y) ** 2)
    right_nose_dist = math.sqrt((right_nose_x - mid_nose_x) ** 2 + (right_nose_y - mid_nose_y) ** 2)
    # Nose-midpoint deviation, pairwise:
    nose_dev = math.sqrt((left_nose_dist - right_nose_dist) ** 2)

    left_skull_x = landmarks.part(0).x
    left_skull_y = landmarks.part(0).y
    right_skull_x = landmarks.part(16).x
    right_skull_y = landmarks.part(16).y
    mid_skull_x = landmarks.part(28).x
    mid_skull_y = landmarks.part(28).y
    left_skull_dist = math.sqrt((left_skull_x - mid_skull_x) ** 2 + (left_skull_y - mid_skull_y) ** 2)
    right_skull_dist = math.sqrt((right_skull_x - mid_skull_x) ** 2 + (right_skull_y - mid_skull_y) ** 2)
    # Skull-midpoint deviation, pairwise:
    skull_dev = math.sqrt((left_skull_dist - right_skull_dist) ** 2)

    left_eye_corner_out_x = landmarks.part(36).x
    left_eye_corner_out_y = landmarks.part(36).y
    left_eye_corner_in_x = landmarks.part(39).x
    left_eye_corner_in_y = landmarks.part(39).y
    right_eye_corner_out_x = landmarks.part(45).x
    right_eye_corner_out_y = landmarks.part(45).y
    right_eye_corner_in_x = landmarks.part(42).x
    right_eye_corner_in_y = landmarks.part(42).y
    # Calculating the size of an eye cut:
    left_eye_size = math.sqrt(
        (left_eye_corner_out_x - left_eye_corner_in_x) ** 2 + (left_eye_corner_out_y - left_eye_corner_in_y) ** 2)
    right_eye_size = math.sqrt(
        (right_eye_corner_out_x - right_eye_corner_in_x) ** 2 + (right_eye_corner_out_y - right_eye_corner_in_y) ** 2)
    # Eye size difference, pairwise:
    eye_size_diff = math.sqrt((left_eye_size - right_eye_size) ** 2)

    left_brow_mid_x = landmarks.part(19).x
    left_brow_mid_y = landmarks.part(19).y
    right_brow_mid_x = landmarks.part(24).x
    right_brow_mid_y = landmarks.part(24).y
    left_eye_up_x = landmarks.part(38).x
    left_eye_up_y = landmarks.part(38).y
    right_eye_up_x = landmarks.part(44).x
    right_eye_up_y = landmarks.part(44).y
    # Calculating deeb -  distance between eye and eyebrow
    left_deeb = math.sqrt((left_brow_mid_x - left_eye_up_x) ** 2 + (left_brow_mid_y - left_eye_up_y) ** 2)
    right_deeb = math.sqrt((right_brow_mid_x - right_eye_up_x) ** 2 + (right_brow_mid_y - right_eye_up_y) ** 2)
    # Deeb deviation, pairwise:
    deeb_dev = math.sqrt((left_deeb - right_deeb) ** 2)

    # Overall symmetry index - Sum of pairwise deviations of face feature points and eye cut difference:
    # Since sum() takes 2 arg at most, paired sums:
    sum1 = sum([jaw_dev, lip_dev])
    sum2 = sum([nose_dev, skull_dev])
    sum3 = sum([deeb_dev, eye_size_diff])

    sum4 = sum([sum1, sum2])
    SymmIndex = sum([sum4, sum3])

    print(SymmIndex)

求助！我卡住了

我想我应该在让 python 读取文件夹时部署一个循环，但我无法思考在哪里以及如何使用它来进一步输出数字数组。