How to Build a 200-Line Python Script for Automatic Face Swapping

Introduction

This guide shows how to write a short (about 200 lines) Python script that automatically replaces the face in one image with the face from another image.

Step 1: Detect facial landmarks with dlib

The script uses dlib’s Python bindings to locate 68 facial landmarks. These points provide the geometry needed for subsequent alignment.

PREDICTOR_PATH = "/home/matt/dlib-18.16/shape_predictor_68_face_landmarks.dat"

detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor(PREDICTOR_PATH)

def get_landmarks(im):
    rects = detector(im, 1)
    if len(rects) > 1:
        raise TooManyFaces
    if len(rects) == 0:
        raise NoFaces
    return numpy.matrix([[p.x, p.y] for p in predictor(im, rects[0]).parts()])

Step 2: Align faces using Procrustes analysis

With two landmark matrices (one for each image) we compute a similarity transform (scale s, rotation R, translation T) that best aligns the first set to the second.

def transformation_from_points(points1, points2):
    points1 = points1.astype(numpy.float64)
    points2 = points2.astype(numpy.float64)
    c1 = numpy.mean(points1, axis=0)
    c2 = numpy.mean(points2, axis=0)
    points1 -= c1
    points2 -= c2
    s1 = numpy.std(points1)
    s2 = numpy.std(points2)
    points1 /= s1
    points2 /= s2
    U, S, Vt = numpy.linalg.svd(points1.T * points2)
    R = (U * Vt).T
    return numpy.vstack([
        numpy.hstack(((s2 / s1) * R, c2.T - (s2 / s1) * R * c1.T)),
        numpy.matrix([0., 0., 1.])
    ])

Step 3: Correct the colour of the second image

Differences in lighting and skin tone are mitigated by adjusting the colour of the second image based on a Gaussian‑blurred version of both images.

COLOUR_CORRECT_BLUR_FRAC = 0.6
LEFT_EYE_POINTS = list(range(42, 48))
RIGHT_EYE_POINTS = list(range(36, 42))

def correct_colours(im1, im2, landmarks1):
    blur_amount = COLOUR_CORRECT_BLUR_FRAC * numpy.linalg.norm(
        numpy.mean(landmarks1[LEFT_EYE_POINTS], axis=0) -
        numpy.mean(landmarks1[RIGHT_EYE_POINTS], axis=0))
    blur_amount = int(blur_amount)
    if blur_amount % 2 == 0:
        blur_amount += 1
    im1_blur = cv2.GaussianBlur(im1, (blur_amount, blur_amount), 0)
    im2_blur = cv2.GaussianBlur(im2, (blur_amount, blur_amount), 0)
    im2_blur += 128 * (im2_blur <= 1.0)
    return (im2.astype(numpy.float64) * im1_blur.astype(numpy.float64) /
            im2_blur.astype(numpy.float64))

Step 4: Blend the second face onto the first

A mask is created that defines which pixels come from the second image and which from the first. The mask is feathered to avoid harsh edges, then the two images are combined.

LEFT_EYE_POINTS = list(range(42, 48))
RIGHT_EYE_POINTS = list(range(36, 42))
LEFT_BROW_POINTS = list(range(22, 27))
RIGHT_BROW_POINTS = list(range(17, 22))
NOSE_POINTS = list(range(27, 35))
MOUTH_POINTS = list(range(48, 61))
OVERLAY_POINTS = [
    LEFT_EYE_POINTS + RIGHT_EYE_POINTS + LEFT_BROW_POINTS + RIGHT_BROW_POINTS,
    NOSE_POINTS + MOUTH_POINTS,
]
FEATHER_AMOUNT = 11

def draw_convex_hull(im, points, color):
    points = cv2.convexHull(points)
    cv2.fillConvexPoly(im, points, color=color)

def get_face_mask(im, landmarks):
    im = numpy.zeros(im.shape[:2], dtype=numpy.float64)
    for group in OVERLAY_POINTS:
        draw_convex_hull(im, landmarks[group], color=1)
    im = numpy.array([im, im, im]).transpose((1, 2, 0))
    im = (cv2.GaussianBlur(im, (FEATHER_AMOUNT, FEATHER_AMOUNT), 0) > 0) * 1.0
    im = cv2.GaussianBlur(im, (FEATHER_AMOUNT, FEATHER_AMOUNT), 0)
    return im

mask = get_face_mask(im2, landmarks2)
warped_mask = warp_im(mask, M, im1.shape)
combined_mask = numpy.max([get_face_mask(im1, landmarks1), warped_mask], axis=0)
output_im = im1 * (1.0 - combined_mask) + warped_corrected_im2 * combined_mask

Results

The final image shows the face from the second picture seamlessly blended onto the first, with aligned geometry and adjusted colour.