def forward_pass(img, session, images_placeholder, phase_train_placeholder, embeddings, image_size):
# If there is a human face
if img is not None:
# Normalize the pixel values of the image for noise reduction for better accuracy and resize to desired size
image = load_img(
img=img, do_random_crop=False, do_random_flip=False,
do_prewhiten=True, image_size=image_size
)
# Run forward pass on FaceNet model to calculate embedding
feed_dict = {images_placeholder: image, phase_train_placeholder: False}
embedding = session.run(embeddings, feed_dict=feed_dict)
return embedding
else:
return None
def recognize_face(sess, pnet, rnet, onet, feature_array):
# Get input and output tensors
images_placeholder = sess.graph.get_tensor_by_name("input:0")
images_placeholder = tf.image.resize_images(images_placeholder, (160, 160))
embeddings = sess.graph.get_tensor_by_name("embeddings:0")
phase_train_placeholder = sess.graph.get_tensor_by_name("phase_train:0")
image_size = args.image_size
embedding_size = embeddings.get_shape()[1]
cap = cv2.VideoCapture(imagePath)
# Default resolutions of the frame are obtained.The default resolutions are system dependent.
# We convert the resolutions from float to integer.
frame_width = int(cap.get(3))
frame_height = int(cap.get(4))
fps = cap.get(cv2.CAP_PROP_FPS)
#out = cv2.VideoWriter(imagePathw,cv2.VideoWriter_fourcc('M','J','P','G'), 10, (frame_width,frame_height))
out = cv2.VideoWriter(imagePathw,
cv2.VideoWriter_fourcc('M', 'J', 'P',
'G'), fps, (1920, 1080))
while (True):
ret, gray = cap.read()
#gray = cv2.cvtColor(frame, 0)
gray = cv2.resize(gray, (1920, 1080))
if ret == True:
#print(gray.size)
#faces = face_cascade.detectMultiScale(gray, 1.3, 5)
response, faces, bboxs = align_face(gray, pnet, rnet, onet)
#print(response)
if (response == True):
for i, image in enumerate(faces):
bb = bboxs[i]
images = load_img(image, False, False, image_size)
feed_dict = {
images_placeholder: images,
phase_train_placeholder: False
}
#feature_vector = sess.run(embeddings, feed_dict=feed_dict)
feature_vector = sess.run(embeddings, feed_dict)
result, diff = identify_person(feature_vector,
feature_array, 5)
#print(result.split("/")[2])
W = int(bb[2] - bb[0]) // 2
H = int(bb[3] - bb[1]) // 2
if (result.split("/")[0] == 'unknown'):
#cv2.rectangle(gray,(bb[0],bb[1]),(bb[2],bb[3]),(255,0,0),2)
#cv2.rectangle(gray,(bb[0]-1,bb[1]),(bb[2]+1,bb[1]-18),(255,0,0),cv2.FILLED)
#cv2.rectangle(gray,(bb[0]-1,bb[3]),(bb[2]+1,bb[3]+18),(255,0,0),cv2.FILLED)
#cv2.putText(gray, result.split("/")[0] + " " + str('%.2f' % diff),(bb[0],bb[1]-5), cv2.FONT_HERSHEY_SIMPLEX,0.5,(255,255,191),1,cv2.LINE_AA)
#cv2.putText(gray, str(diff),(bb[0],bb[3]+10), cv2.FONT_HERSHEY_SIMPLEX,0.5,(255,255,255),1,cv2.LINE_AA)
putTextName(gray, bb,
result.split("/")[0], str(diff),
(0, 0, 255))
print("\x1b[1;31;40mUnknown " + "\x1b[0m" +
result.split("/")[1])
else:
#cv2.rectangle(gray,(bb[0],bb[1]),(bb[2],bb[3]),(0,255,0),2)
#cv2.rectangle(gray,(bb[0]-1,bb[1]),(bb[2]+1,bb[1]-18),(0,255,0),cv2.FILLED)
#cv2.rectangle(gray,(bb[0]-1,bb[3]),(bb[2]+1,bb[3]+18),(0,255,0),cv2.FILLED)
#cv2.putText(gray, result.split("/")[0] + " " + str('%.2f' % diff),(bb[0],bb[1]-5), cv2.FONT_HERSHEY_SIMPLEX,0.5,(0,0,0),1,cv2.LINE_AA)
#cv2.putText(gray, str(diff),(bb[0],bb[3]+10), cv2.FONT_HERSHEY_SIMPLEX,0.5,(0,0,0),1,cv2.LINE_AA)
putTextName(gray, bb,
result.split("/")[0], str(diff),
(255, 255, 127))
print("\x1b[1;32;40mMatched " + "\x1b[0m" +
result.split("/")[1])
#gray = cv2.resize(gray, (1280, 720))
out.write(gray)
def recognize_face(sess, pnet, rnet, onet, feature_array):
# Get input and output tensors
images_placeholder = sess.graph.get_tensor_by_name("input:0")
images_placeholder = tf.image.resize_images(images_placeholder, (160, 160))
embeddings = sess.graph.get_tensor_by_name("embeddings:0")
phase_train_placeholder = sess.graph.get_tensor_by_name("phase_train:0")
image_size = args.image_size
embedding_size = embeddings.get_shape()[1]
#cap = cv2.VideoCapture(-1)
#while(True):
#ret, frame = cap.read()
#gray = cv2.cvtColor(frame, 0)
img = misc.imread(imagePath)
#gray = cv2.cvtColor(img, 0)
gray = img
#if cv2.waitKey(1) & 0xFF == ord('q'):
# cap.release()
# cv2.destroyAllWindows()
# break
if (gray.size > 0):
print(gray.size)
#faces = face_cascade.detectMultiScale(gray, 1.3, 5)
response, faces, bboxs = align_face(gray, pnet, rnet, onet)
print(response)
if (response == True):
for i, image in enumerate(faces):
bb = bboxs[i]
images = load_img(image, False, False, image_size)
feed_dict = {
images_placeholder: images,
phase_train_placeholder: False
}
#feature_vector = sess.run(embeddings, feed_dict=feed_dict)
feature_vector = sess.run(embeddings, feed_dict)
result, diff = identify_person(feature_vector, feature_array,
5)
#print(result.split("/")[2])
W = int(bb[2] - bb[0]) // 2
H = int(bb[3] - bb[1]) // 2
if (result.split("/")[0] == 'unknown'):
#cv2.rectangle(gray,(bb[0],bb[1]),(bb[2],bb[3]),(255,0,0),2)
#cv2.rectangle(gray,(bb[0]-1,bb[1]),(bb[2]+1,bb[1]-18),(255,0,0),cv2.FILLED)
#cv2.rectangle(gray,(bb[0]-1,bb[3]),(bb[2]+1,bb[3]+18),(255,0,0),cv2.FILLED)
#cv2.putText(gray, result.split("/")[0] + " " + str('%.2f' % diff),(bb[0],bb[1]-5), cv2.FONT_HERSHEY_SIMPLEX,0.5,(255,255,191),1,cv2.LINE_AA)
#cv2.putText(gray, str(diff),(bb[0],bb[3]+10), cv2.FONT_HERSHEY_SIMPLEX,0.5,(255,255,255),1,cv2.LINE_AA)
putTextName(gray, bb,
result.split("/")[0], str(diff), (255, 0, 0))
print("\x1b[1;31;40mUnknown " + "\x1b[0m" +
result.split("/")[1])
else:
#cv2.rectangle(gray,(bb[0],bb[1]),(bb[2],bb[3]),(0,255,0),2)
#cv2.rectangle(gray,(bb[0]-1,bb[1]),(bb[2]+1,bb[1]-18),(0,255,0),cv2.FILLED)
#cv2.rectangle(gray,(bb[0]-1,bb[3]),(bb[2]+1,bb[3]+18),(0,255,0),cv2.FILLED)
#cv2.putText(gray, result.split("/")[0] + " " + str('%.2f' % diff),(bb[0],bb[1]-5), cv2.FONT_HERSHEY_SIMPLEX,0.5,(0,0,0),1,cv2.LINE_AA)
#cv2.putText(gray, str(diff),(bb[0],bb[3]+10), cv2.FONT_HERSHEY_SIMPLEX,0.5,(0,0,0),1,cv2.LINE_AA)
putTextName(gray, bb,
result.split("/")[0], str(diff),
(127, 255, 255))
print("\x1b[1;32;40mMatched " + "\x1b[0m" +
result.split("/")[1])
misc.imsave(imagePathw, gray)
def main(args):
with tf.Graph().as_default():
with tf.Session() as sess:
# Get the paths for the corresponding images
vinayak = [
'datasets/kar_Vin_aligned/vinayak/' + f
for f in os.listdir('datasets/kar_Vin_aligned/vinayak')
]
karthik = [
'datasets/kar_Vin_aligned/karthik/' + f
for f in os.listdir('datasets/kar_Vin_aligned/karthik')
]
ashish = [
'datasets/kar_Vin_aligned/Ashish/' + f
for f in os.listdir('datasets/kar_Vin_aligned/Ashish')
]
saurabh = [
'datasets/kar_Vin_aligned/Saurabh/' + f
for f in os.listdir('datasets/kar_Vin_aligned/Saurabh')
]
hari = [
'datasets/kar_Vin_aligned/Hari/' + f
for f in os.listdir('datasets/kar_Vin_aligned/Hari')
]
paths = vinayak + karthik + ashish + saurabh + hari
#np.save("images.npy",paths)
# Load the model
facenet.load_model(args.model)
# Get input and output tensors
images_placeholder = tf.get_default_graph().get_tensor_by_name(
"input:0")
images_placeholder = tf.image.resize_images(
images_placeholder, (160, 160))
embeddings = tf.get_default_graph().get_tensor_by_name(
"embeddings:0")
phase_train_placeholder = tf.get_default_graph(
).get_tensor_by_name("phase_train:0")
image_size = args.image_size
embedding_size = embeddings.get_shape()[1]
extracted_dict = {}
# Run forward pass to calculate embeddings
for i, filename in enumerate(paths):
images = facenet.load_img(filename, False, False, image_size)
feed_dict = {
images_placeholder: images,
phase_train_placeholder: False
}
feature_vector = sess.run(embeddings, feed_dict=feed_dict)
extracted_dict[filename] = feature_vector
if (i % 100 == 0):
print("completed", i, " images")
with open('extracted_dict.pickle', 'wb') as f:
pickle.dump(extracted_dict, f)
def recognize_face(sess, pnet, rnet, onet, feature_array):
# Get input and output tensors
images_placeholder = sess.graph.get_tensor_by_name("input:0")
images_placeholder = tf.image.resize_images(images_placeholder, (160, 160))
embeddings = sess.graph.get_tensor_by_name("embeddings:0")
phase_train_placeholder = sess.graph.get_tensor_by_name("phase_train:0")
image_size = args.image_size
embedding_size = embeddings.get_shape()[1]
#cap = cv2.VideoCapture(-1)
cap = cv2.VideoCapture(0)
while (True):
ret, frame = cap.read()
#gray = cv2.cvtColor(frame, 0)
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# To close the window
if cv2.waitKey(1) & 0xFF == ord('q'):
cap.release()
cv2.destroyAllWindows()
break
if (gray.size > 0):
print(gray.size)
response, faces, bboxs = align_face(gray, pnet, rnet, onet)
print(response)
if (response == True):
for i, image in enumerate(faces):
bb = bboxs[i]
images = load_img(image, False, False, image_size)
feed_dict = {
images_placeholder: images,
phase_train_placeholder: False
}
feature_vector = sess.run(embeddings, feed_dict=feed_dict)
result, accuracy = identify_person(feature_vector,
feature_array, 8)
#print(type(result))
#print(len(result.split("/")))
print(result.split("\\")[7])
print(accuracy)
if accuracy < 0.9:
cv2.rectangle(gray, (bb[0], bb[1]), (bb[2], bb[3]),
(255, 255, 255), 2)
W = int(bb[2] - bb[0]) // 2
H = int(bb[3] - bb[1]) // 2
#cv2.putText(gray,"Hello "+result.split['/'][0],(bb[0]+W-(W//2),bb[1]-7), cv2.FONT_HERSHEY_SIMPLEX,0.5,(255,255,255),1,cv2.LINE_AA)
cv2.putText(
gray,
"Hello " + result.split("\\")[7].split('.')[0],
(bb[0] + W - (W // 2), bb[1] - 7),
cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 255), 1,
cv2.LINE_AA)
else:
cv2.rectangle(gray, (bb[0], bb[1]), (bb[2], bb[3]),
(255, 255, 255), 2)
W = int(bb[2] - bb[0]) // 2
H = int(bb[3] - bb[1]) // 2
cv2.putText(gray, "WHO ARE YOU ?",
(bb[0] + W - (W // 2), bb[1] - 7),
cv2.FONT_HERSHEY_SIMPLEX, 0.5,
(255, 255, 255), 1, cv2.LINE_AA)
del feature_vector
cv2.imshow('img', gray)
else:
continue