def face_detection(image, resize, video=False):
if resize:
image = imutils.resize(image, width=500)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# detect faces in the grayscale image
rects = detector(gray, 1)
# loop over the face detections
for (i, rect) in enumerate(rects):
# determine the facial landmarks for the face region, then
# convert the facial landmark (x, y)-coordinates to a NumPy
# array
shape = predictor(gray, rect)
shape = face_utils.shape_to_np(shape)
# convert dlib's rectangle to a OpenCV-style bounding box
# [i.e., (x, y, w, h)], then draw the face bounding box
(x, y, w, h) = face_utils.rect_to_bb(rect)
# cv2.rectangle(image, (x, y), (x + w, y + h), (0, 255, 0), 2)
# show the face number
# cv2.putText(image, "Face #{}".format(i + 1), (x - 10, y - 10),
# cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)
# loop over the (x, y)-coordinates for the facial landmarks
# and draw them on the image
if video:
for (x, y) in shape:
cv2.circle(image, (x, y), 1, (255, 0, 0), -1)
else:
image = visualize_facial_landmarks(image, shape)
return image
face_img = np.zeros((64, 96)).astype(np.uint8)
left_img = np.zeros((64, 96)).astype(np.uint8)
rigt_img = np.zeros((64, 96)).astype(np.uint8)
cur_status = None
cur_direction = None
cur_spacekey = None
for (ii, rect_s) in enumerate(rects_small):
tmp = np.array([rect_s.left(), rect_s.top(), rect_s.right(), rect_s.bottom()]) / scale
tmp = tmp.astype(np.long)
# get face rect
rect = dlib.rectangle(tmp[0], tmp[1], tmp[2], tmp[3])
(x, y, w, h) = face_utils.rect_to_bb(rect)
cv2.rectangle(frame_small, (int(x*scale), int(y*scale)), (int((x+w)*scale), int((y+h)*scale)), (0, 255, 0), 2)
# get face landmarks
shape = predictor(frame, rect)
shape = face_utils.shape_to_np(shape)
cur_status = face_utils.get_mouth_status(shape)
for i, (x, y) in enumerate(shape):
cv2.circle(frame_small, (int(x*scale), int(y*scale)), 1, (0, 0, 255), -1)
# eye gaze estimation
face_img, left_img, rigt_img, eye_lm, fc_c_world = \
pre_eye.WarpNCrop(frame[:,:,::-1], shape, inv_cameraMat, cam_new)
y_result, eye_tensor, face_tensor = sess.run([y_conv, h_trans, face_h_trans], feed_dict={
# show the original input image and detect faces in the grayscale image
if DEBUG:
cv2.imshow("Input", image)
cv2.waitKey(1)
[rects, scores, tmp] = detector.run(gray, 2)
if len(rects) > 0: # check if exist some detection
ix = scores.index(max(scores))
rect = rects[ix]
# extract the ROI of the *original* face, then align the face
# using facial landmarks
(x, y, w, h) = rect_to_bb(rect)
faceAligned, pose = fa.align(image, gray, rect)
# display the output images
if DEBUG:
faceOrig = imutils.resize(image[y:y + h, x:x + w], width=256)
cv2.imshow("Original", faceOrig)
cv2.imshow("Aligned", faceAligned)
cv2.waitKey(1)
if pose:
path, file = os.path.split(image_path)
if path != last_path:
image_path_out = path.replace(dataset_path, output_path)
last_path = path