def __init__(
self,
desiredFaceWidth=160,
predictor_path='models/shape_predictor_68_face_landmarks.dat',
):
self.predictor = dlib.shape_predictor(predictor_path)
self.fa = FaceAligner(self.predictor,
desiredFaceWidth=desiredFaceWidth,
desiredLeftEye=(0.32, 0.35))
self.encoder = Encoder()
def detect_face(img):
sp = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')
fa = FaceAligner(sp)
#img = imutils.resize(img, width=500)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
detector = dlib.get_frontal_face_detector()
rects = detector(gray, 2)
num_faces = len(rects)
if num_faces == 0:
return None
faceAligned = fa.align(img, gray, rects[0])
image = cv2.cvtColor(faceAligned, cv2.COLOR_BGR2GRAY)
return image
class EnhancedFacenet:
def __init__(self,desiredFaceWidth=160, predictor_path='models/shape_predictor_68_face_landmarks.dat', ):
self.predictor = dlib.shape_predictor(predictor_path)
self.fa = FaceAligner(self.predictor, desiredFaceWidth=desiredFaceWidth, desiredLeftEye=(0.37, 0.33))
self.encoder = Encoder()
def alignAndEncode(self, img, gray, face_rect):
face = self.fa.align(img, gray, face_rect )
face_rgb = cv2.cvtColor(face, cv2.COLOR_BGR2RGB)
return face, self.encoder.generate_embedding(face_rgb)
def align_faces(imagePaths: list):
"""
Align faces found in imagePaths,
overwrite original with the aligned face
- imagePaths: list of paths of images to align
"""
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor(
"models/shape_predictor_68_face_landmarks.dat")
fa = FaceAligner(predictor, desiredFaceHeight=256)
for i, imagePath in enumerate(imagePaths):
print("[INFO] - Aligning face #{}".format(i))
image = cv2.imread(imagePath)
image = imutils.resize(image, width=800)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
rects = detector(gray, 2)
for rect in rects:
image = fa.align(image, gray, rect)
cv2.imwrite(imagePath, image)
def detect_face(img):
sp = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')
fa = FaceAligner(sp)
#img = imutils.resize(img, width=500)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
detector = dlib.get_frontal_face_detector()
rects = detector(gray, 2)
num_faces = len(rects)
if num_faces == 0:
return None
b = 0
l = 0
for rect in rects:
# extract the ROI of the *original* face, then align the face
# using facial landmarks
(x, y, w, h) = face_utils.rect_to_bb(rect)
#print(w," ",h)
if w > b and h > l:
b = w
l = h
faceAligned = fa.align(img, gray, rect)
image = cv2.cvtColor(faceAligned, cv2.COLOR_BGR2GRAY)
return image
import dlib
import cv2
import os
import glob
from skimage import io
import imutils
from imutils import face_utils
import cv2
import os
import numpy as np
from imutils.face_utils.facealigner import FaceAligner
sp = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')
fa = FaceAligner(sp)
# load the input image, resize it, and convert it to grayscale
image = cv2.imread('cr71.png')
image = imutils.resize(image, width=800)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# show the original input image and detect faces in the grayscale
# image
cv2.imshow("Input", image)
detector = dlib.get_frontal_face_detector()
rects = detector(gray, 2)
for rect in rects:
# extract the ROI of the *original* face, then align the face
# using facial landmarks
(x, y, w, h) = face_utils.rect_to_bb(rect)
faceOrig = imutils.resize(image[y:y + h, x:x + w], width=256)
faceAligned = fa.align(image, gray, rect)