小200行Python代碼做了一個換臉程序
簡介
在這篇文章中我將介紹如何寫一個簡短(200行)的 Python 腳本,來自動地將一幅圖片的臉替換為另一幅圖片的臉。
這個過程分四步:
-
檢測臉部標(biāo)記。
-
旋轉(zhuǎn)、縮放、平移和第二張圖片,以配合***步。
-
調(diào)整第二張圖片的色彩平衡,以適配***張圖片。
-
把第二張圖像的特性混合在***張圖像中。
1.使用 dlib 提取面部標(biāo)記
該腳本使用 dlib 的 Python 綁定來提取面部標(biāo)記:
Dlib 實現(xiàn)了 Vahid Kazemi 和 Josephine Sullivan 的《使用回歸樹一毫秒臉部對準(zhǔn)》論文中的算法。算法本身非常復(fù)雜,但dlib接口使用起來非常簡單:
- 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()])
get_landmarks()函數(shù)將一個圖像轉(zhuǎn)化成numpy數(shù)組,并返回一個68×2元素矩陣,輸入圖像的每個特征點對應(yīng)每行的一個x,y坐標(biāo)。
特征提取器(predictor)需要一個粗糙的邊界框作為算法輸入,由一個傳統(tǒng)的能返回一個矩形列表的人臉檢測器(detector)提供,其每個矩形列表在圖像中對應(yīng)一個臉。
2.用 Procrustes 分析調(diào)整臉部
現(xiàn)在我們已經(jīng)有了兩個標(biāo)記矩陣,每行有一組坐標(biāo)對應(yīng)一個特定的面部特征(如第30行的坐標(biāo)對應(yīng)于鼻頭)。我們現(xiàn)在要解決如何旋轉(zhuǎn)、翻譯和縮放***個向量,使它們盡可能適配第二個向量的點。一個想法是可以用相同的變換在***個圖像上覆蓋第二個圖像。
將這個問題數(shù)學(xué)化,尋找T,s 和 R,使得下面這個表達(dá)式:
結(jié)果最小,其中R是個2×2正交矩陣,s是標(biāo)量,T是二維向量,pi和qi是上面標(biāo)記矩陣的行。
事實證明,這類問題可以用“常規(guī) Procrustes 分析法”解決:
- 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.])])
代碼實現(xiàn)了這幾步:
1.將輸入矩陣轉(zhuǎn)換為浮點數(shù)。這是后續(xù)操作的基礎(chǔ)。
2.每一個點集減去它的矩心。一旦為點集找到了一個***的縮放和旋轉(zhuǎn)方法,這兩個矩心 c1 和 c2 就可以用來找到完整的解決方案。
3.同樣,每一個點集除以它的標(biāo)準(zhǔn)偏差。這會消除組件縮放偏差的問題。
4.使用奇異值分解計算旋轉(zhuǎn)部分??梢栽诰S基百科上看到關(guān)于解決正交 Procrustes 問題的細(xì)節(jié)。
5.利用仿射變換矩陣返回完整的轉(zhuǎn)化。
其結(jié)果可以插入 OpenCV 的 cv2.warpAffine 函數(shù),將圖像二映射到圖像一:
- def warp_im(im, M, dshape):
- output_im = numpy.zeros(dshape, dtype=im.dtype)
- cv2.warpAffine(im,
- M[:2],
- (dshape[1], dshape[0]),
- dst=output_im,
- borderMode=cv2.BORDER_TRANSPARENT,
- flags=cv2.WARP_INVERSE_MAP)
- return output_im
對齊結(jié)果如下:
3.校正第二張圖像的顏色
如果我們試圖直接覆蓋面部特征,很快會看到這個問題:
這個問題是兩幅圖像之間不同的膚色和光線造成了覆蓋區(qū)域的邊緣不連續(xù)。我們試著修正:
- 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)
- # Avoid divide-by-zero errors.
- im2_blur += 128 * (im2_blur <= 1.0)
- return (im2.astype(numpy.float64) * im1_blur.astype(numpy.float64) /
- im2_blur.astype(numpy.float64))
結(jié)果如下:
此函數(shù)試圖改變 im2 的顏色來適配 im1。它通過用 im2 除以 im2 的高斯模糊值,然后乘以im1的高斯模糊值。這里的想法是用RGB縮放校色,但并不是用所有圖像的整體常數(shù)比例因子,每個像素都有自己的局部比例因子。
用這種方法兩圖像之間光線的差異只能在某種程度上被修正。例如,如果圖像1是從一側(cè)照亮,但圖像2是被均勻照亮的,色彩校正后圖像2也會出現(xiàn)未照亮一側(cè)暗一些的問題。
也就是說,這是一個相當(dāng)簡陋的辦法,而且解決問題的關(guān)鍵是一個適當(dāng)?shù)母咚购撕瘮?shù)大小。如果太小,***個圖像的面部特征將顯示在第二個圖像中。過大,內(nèi)核之外區(qū)域像素被覆蓋,并發(fā)生變色。這里的內(nèi)核用了一個0.6 *的瞳孔距離。
4.把第二張圖像的特征混合在***張圖像中
用一個遮罩來選擇圖像2和圖像1的哪些部分應(yīng)該是最終顯示的圖像:
值為1(顯示為白色)的地方為圖像2應(yīng)該顯示出的區(qū)域,值為0(顯示為黑色)的地方為圖像1應(yīng)該顯示出的區(qū)域。值在0和1之間為圖像1和圖像2的混合區(qū)域。
這是生成上圖的代碼:
- 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)
我們把上述過程分解:
-
get_face_mask()的定義是為一張圖像和一個標(biāo)記矩陣生成一個遮罩,它畫出了兩個白色的凸多邊形:一個是眼睛周圍的區(qū)域,一個是鼻子和嘴部周圍的區(qū)域。之后它由11個像素向遮罩的邊緣外部羽化擴(kuò)展,可以幫助隱藏任何不連續(xù)的區(qū)域。
-
這樣一個遮罩同時為這兩個圖像生成,使用與步驟2中相同的轉(zhuǎn)換,可以使圖像2的遮罩轉(zhuǎn)化為圖像1的坐標(biāo)空間。
-
之后,通過一個element-wise***值,這兩個遮罩結(jié)合成一個。結(jié)合這兩個遮罩是為了確保圖像1被掩蓋,而顯現(xiàn)出圖像2的特性。
***,使用遮罩得到最終的圖像:
- output_im = im1 * (1.0 - combined_mask) + warped_corrected_im2 * combined_mask
完整代碼(link):
- import cv2
- import dlib
- import numpy
- import sys
- PREDICTOR_PATH = "/home/matt/dlib-18.16/shape_predictor_68_face_landmarks.dat"
- SCALE_FACTOR = 1
- FEATHER_AMOUNT = 11
- FACE_POINTS = list(range(17, 68))
- MOUTH_POINTS = list(range(48, 61))
- RIGHT_BROW_POINTS = list(range(17, 22))
- LEFT_BROW_POINTS = list(range(22, 27))
- RIGHT_EYE_POINTS = list(range(36, 42))
- LEFT_EYE_POINTS = list(range(42, 48))
- NOSE_POINTS = list(range(27, 35))
- JAW_POINTS = list(range(0, 17))
- # Points used to line up the images.
- ALIGN_POINTS = (LEFT_BROW_POINTS + RIGHT_EYE_POINTS + LEFT_EYE_POINTS +
- RIGHT_BROW_POINTS + NOSE_POINTS + MOUTH_POINTS)
- # Points from the second image to overlay on the first. The convex hull of each
- # element will be overlaid.
- OVERLAY_POINTS = [
- LEFT_EYE_POINTS + RIGHT_EYE_POINTS + LEFT_BROW_POINTS + RIGHT_BROW_POINTS,
- NOSE_POINTS + MOUTH_POINTS,
- ]
- # Amount of blur to use during colour correction, as a fraction of the
- # pupillary distance.
- COLOUR_CORRECT_BLUR_FRAC = 0.6
- detector = dlib.get_frontal_face_detector()
- predictor = dlib.shape_predictor(PREDICTOR_PATH)
- class TooManyFaces(Exception):
- pass
- class NoFaces(Exception):
- pass
- 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()])
- def annotate_landmarks(im, landmarks):
- im = im.copy()
- for idx, point in enumerate(landmarks):
- pos = (point[0, 0], point[0, 1])
- cv2.putText(im, str(idx), pos,
- fontFace=cv2.FONT_HERSHEY_SCRIPT_SIMPLEX,
- fontScale=0.4,
- color=(0, 0, 255))
- cv2.circle(im, pos, 3, color=(0, 255, 255))
- return im
- 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
- def transformation_from_points(points1, points2):
- """
- Return an affine transformation [s * R | T] such that:
- sum ||s*R*p1,i + T - p2,i||^2
- is minimized.
- """
- # Solve the procrustes problem by subtracting centroids, scaling by the
- # standard deviation, and then using the SVD to calculate the rotation. See
- # the following for more details:
- # https://en.wikipedia.org/wiki/Orthogonal_Procrustes_problem
- 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)
- # The R we seek is in fact the transpose of the one given by U * Vt. This
- # is because the above formulation assumes the matrix goes on the right
- # (with row vectors) where as our solution requires the matrix to be on the
- # left (with column vectors).
- R = (U * Vt).T
- return numpy.vstack([numpy.hstack(((s2 / s1) * R,
- c2.T - (s2 / s1) * R * c1.T)),
- numpy.matrix([0., 0., 1.])])
- def read_im_and_landmarks(fname):
- im = cv2.imread(fname, cv2.IMREAD_COLOR)
- im = cv2.resize(im, (im.shape[1] * SCALE_FACTOR,
- im.shape[0] * SCALE_FACTOR))
- s = get_landmarks(im)
- return im, s
- def warp_im(im, M, dshape):
- output_im = numpy.zeros(dshape, dtype=im.dtype)
- cv2.warpAffine(im,
- M[:2],
- (dshape[1], dshape[0]),
- dst=output_im,
- borderMode=cv2.BORDER_TRANSPARENT,
- flags=cv2.WARP_INVERSE_MAP)
- return output_im
- 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)
- # Avoid divide-by-zero errors.
- im2_blur += 128 * (im2_blur <= 1.0)
- return (im2.astype(numpy.float64) * im1_blur.astype(numpy.float64) /
- im2_blur.astype(numpy.float64))
- im1, landmarks1 = read_im_and_landmarks(sys.argv[1])
- im2, landmarks2 = read_im_and_landmarks(sys.argv[2])
- M = transformation_from_points(landmarks1[ALIGN_POINTS],
- landmarks2[ALIGN_POINTS])
- 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)
- warped_im2 = warp_im(im2, M, im1.shape)
- warped_corrected_im2 = correct_colours(im1, warped_im2, landmarks1)
- output_im = im1 * (1.0 - combined_mask) + warped_corrected_im2 * combined_mask
- cv2.imwrite('output.jpg', output_im)