def fun_util():
emojis = get_emojis()
disp_probab, disp_class = 0, 0
while True:
img = cam.read()[1]
img = cv2.flip(img, 1)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = detector(gray)
if len(faces) > 0:
for i, face in enumerate(faces):
shape_68 = shape_predictor_68(img, face)
shape = face_utils.shape_to_np(shape_68)
mask = create_mask(shape, img)
masked = cv2.bitwise_and(gray, mask)
maskAligned = fa.align(mask, gray, face)
faceAligned = fa.align(masked, gray, face)
(x0, y0, x1, y1) = get_bounding_rect(maskAligned)
faceAligned = faceAligned[y0:y1, x0:x1]
faceAligned = cv2.resize(faceAligned, (100, 100))
(x, y, w, h) = face_utils.rect_to_bb(face)
#cv2.rectangle(img, (x, y), (x+w, y+h), (255, 255, 0), 2)
cv2.imshow('faceAligned', faceAligned)
cv2.imshow('face #{}'.format(i), img[y:y + h, x:x + w])
pred_probab, pred_class = keras_predict(cnn_model, faceAligned)
img = blend(img, emojis[pred_class], (x, y, w, h))
cv2.imshow('img', img)
if cv2.waitKey(1) == ord('q'):
break
def recognize():
disp_probab, disp_class = 0, 0
while True:
img = cam.read()[1]
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = detector(gray)
if len(faces) > 0:
face = faces[0]
shape_68 = shape_predictor_68(img, face)
shape = face_utils.shape_to_np(shape_68)
mask = create_mask(shape, img)
masked = cv2.bitwise_and(gray, mask)
(x, y, w, h) = face_utils.rect_to_bb(face)
cv2.rectangle(img, (x, y), (x + w, y + h), (255, 255, 0), 2)
faceAligned = fa.align(masked, gray, face)
cv2.imshow('faceAligned', faceAligned)
pred_probab, pred_class = keras_predict(cnn_model, faceAligned)
if pred_probab > 0.5:
disp_probab = pred_probab
disp_class = pred_class
cv2.putText(img, str(disp_probab), (50, 50),
cv2.FONT_HERSHEY_TRIPLEX, 1.5, (0, 0, 0))
cv2.putText(img, str(disp_class), (50, 100),
cv2.FONT_HERSHEY_TRIPLEX, 1.5, (0, 0, 0))
cv2.imshow('img', img)
if cv2.waitKey(1) == ord('q'):
break
def emojify(model):
emojis = get_emojis()
disp_probab, disp_class = 0, 0
while True:
img = cam.read()[1]
img = cv2.flip(img, 1)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = detector(gray)
if len(faces) > 0:
for i, face in enumerate(faces):
shape_68 = shape_predictor_68(img, face)
shape = face_utils.shape_to_np(shape_68)
mask = create_mask(shape, img)
masked = cv2.bitwise_and(gray, mask)
maskAligned = fa.align(mask, gray, face)
faceAligned = fa.align(masked, gray, face)
(x0, y0, x1, y1) = get_bounding_rect(maskAligned)
faceAligned = faceAligned[y0:y1, x0:x1]
cv2.imshow('faceAligned', faceAligned)
(x, y, w, h) = face_utils.rect_to_bb(face)
cv2.imshow('face #{}'.format(i), img[y:y+h, x:x+w])
np_image = load(faceAligned)
pred_probab = model.predict(np_image)
pred_class = np.argmax(pred_probab,axis = 1)
img = blend(img, emojis[int(pred_class)], (x, y, w, h))
cv2.imshow('img', img)
if cv2.waitKey(1) == ord('q'):
break
FACES = 'sorted + ours/'
folders = os.listdir(FACES)
for folder in folders:
folder = folder + '/'
c = 0
for image in os.listdir(FACES + folder):
path = FACES + folder + image
img = cv2.imread(path)
print(image)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = detector(gray)
if len(faces) > 0:
for face in faces:
shape_68 = shape_predictor_68(img, face)
shape = face_utils.shape_to_np(shape_68)
mask = create_mask(shape, img)
masked = cv2.bitwise_and(gray, mask)
maskAligned = fa.align(mask, gray, face)
faceAligned = fa.align(masked, gray, face)
(x0, y0, x1, y1) = get_bounding_rect(maskAligned)
faceAligned = faceAligned[y0:y1, x0:x1]
faceAligned = cv2.resize(faceAligned, (128, 128))
(x, y, w, h) = face_utils.rect_to_bb(face)
cv2.rectangle(img, (x, y), (x + w, y + h), (255, 255, 0), 2)
cv2.imshow('faceAligned', faceAligned)
if not os.path.isdir(DATASET + folder):
os.mkdir(DATASET + folder)
cv2.imwrite(DATASET + folder + image, faceAligned)
c += 1