1

I used dlib and numpy to extract the list of facial keypoints.

Code:

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()])


for f in glob.glob(os.path.join(faces_folder_path, "*")):
    print("Processing file: {}".format(f))
    img = io.imread(f)

    win.clear_overlay()
    win.set_image(img)

    dets = detector(img, 1)
    print("Number of faces detected: {}".format(len(dets)))
    for k, d in enumerate(dets):
        # Get the landmarks/parts for the face in box d.
        shape = predictor(img, d)
        lms = get_landmarks(img)
        print ("Detection {}: Left: {} Top: {} Right: {} Bottom: {}".format(k, d.left(), d.top(), d.right(), d.bottom()))
        print ("Part 0: {}, Part 1: {} ...".format(shape.part(0), shape.part(1)))
        newSection()
        print ("Keypoints:" + (str(lms)))
        # Draw the face landmarks on the screen.
        win.add_overlay(shape)




    win.add_overlay(dets)
    dlib.hit_enter_to_continue()

Result: And as you can see, it works fine. However, what I need isn't the blue alignment lines, but for there to be a numbered correlation between the keypoints and the image, like this. How can I overlay the numpy matrix with the image points:

numpy.matrix([[p.x, p.y] for p in predictor(im, rects[0]).parts()])

on the opencv image? I attempted it myself: 

for idx, point in enumerate(lms):
     pos = (point[0, 0], point[0, 1])
     cv2.putText(imB, 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))

     WIDTH = 1000
     HEIGHT = 1000

     cv2.namedWindow('image', cv2.WINDOW_NORMAL)
     cv2.imshow('image', imB)
     cv2.resizeWindow('image', WIDTH, HEIGHT)

But it just gives me a blank grew box with a red box in it.

Rich
  • 1,103
  • 1
  • 15
  • 36
  • Is `imB` your original image or it is a new image? Also, why do you put the circle on `im`and not in `imB`? Since you are not showing `im` there is no point to it right? – JrBenito Jun 07 '16 at 21:09
  • Ok, I just changed `cv2.circle(im` to `cv2.circle(imB`. It's still a grey box of blankness. – Rich Jun 07 '16 at 21:28
  • and where does imB come from? On your first piece of code your image is stored as `img` not `imB`. Are you loading your image with opencv to imB? – JrBenito Jun 07 '16 at 21:30
  • Well, some of the code came from here: https://github.com/matthewearl/faceswap/blob/master/faceswap.py – Rich Jun 07 '16 at 21:45
  • I kind of just mashed them together, sorry. – Rich Jun 07 '16 at 21:46
  • If both pieces of code are together in your full script, then you should use variable `img` on `cv2.xxx` functions. If your are passing it as `imB`to another function, than try just show (`cv.imshow('image', imB)`) without any processing on the second piece of code (do not use `cv2.put...`). – JrBenito Jun 07 '16 at 21:55
  • Let us [continue this discussion in chat](http://chat.stackoverflow.com/rooms/114072/discussion-between-jrbenito-and-rich). – JrBenito Jun 07 '16 at 21:55

0 Answers0