def visualize_tracks(tracks, image_root, seq, args):
frames = np.unique(tracks[:, 0])
start_frame = min(frames)
save_dir = osp.join(image_root, "results", "tracking_noid", args.detector,
seq)
for f in range(1, int(max(frames))):
data = tracks[tracks[:, 0] == int(f)]
image_path = osp.join(image_root, "images", seq,
"{:06d}.jpg".format(int(f)))
frame = cv2.imread(image_path)
if len(data) > 0:
for box_data in data:
identity = box_data[1]
face_box = box_data[2:6].astype(int)
body_box = box_data[6:].astype(int)
# frame = draw_box_name(face_box, "{:02d}".format(int(identity)), frame)
# frame = draw_box_name(body_box, "{:02d}".format(int(identity)), frame)
frame = draw_box_name(face_box, "", frame)
frame = draw_box_name(body_box, "", frame)
save_path = osp.join(save_dir, "{:06d}_smoothed.jpg".format(int(f)))
cv2.imwrite(save_path, frame)
os.system(
"ffmpeg -framerate 20 -start_number {} -i {}/%06d_smoothed.jpg {}/{}_smoothed.mp4"
.format(start_frame, save_dir, save_dir, seq))
pass
def main(self):
while cap.isOpened():
isSuccess,frame = cap.read()
if isSuccess:
try:
# image = Image.fromarray(frame[...,::-1]) #bgr to rgb
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit, conf.min_face_size)
bboxes = bboxes[:,:-1] #shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1,-1,1,1] # personal choice
results, score = learner.infer(conf, faces, targets, args.tta)
# print(score[0])
for idx,bbox in enumerate(bboxes):
if args.score:
frame = draw_box_name(bbox, names[results[idx] + 1] + '_{:.2f}'.format(score[idx]), frame)
else:
if float('{:.2f}'.format(score[idx])) > .98:
name = names[0]
else:
name = names[results[idx]+1]
frame = draw_box_name(bbox, names[results[idx] + 1], frame)
except:
pass
ret, jpeg = cv2.imencode('.jpg', frame)
return jpeg.tostring()
# cv2.imshow('Arc Face Recognizer', frame)
if cv2.waitKey(1)&0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()
def inference(self,conf,img):
mtcnn = MTCNN()
learner = face_learner(conf,True)
learner.load_state(conf,'final.pth',True,True)
learner.model.eval()
targets, names = load_facebank(conf)
image = Image.open(img)
frame = cv2.imread(img,cv2.IMREAD_COLOR)
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
try:
bboxes, faces = mtcnn.align_multi(image, conf.face_limit, conf.min_face_size)
bboxes = bboxes[:,:-1] #shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1,-1,1,1] # personal choice
results, score = learner.infer(conf, faces, targets, False)
name = names[results[0]+1]
frame = draw_box_name(bboxes[0], name, frame)
except Exception as ex:
name = "Can't detect face."
h, w, c = frame.shape
bbox = [int(h*0.5),int(w*0.5),int(h*0.5),int(w*0.5)]
frame = draw_box_name(bbox, name, frame)
return name, frame
def main():
while cap.isOpened():
isSuccess, frame = cap.read()
match_score = None
name = None
det_image = None
if isSuccess:
try:
# image = Image.fromarray(frame[...,::-1]) #bgr to rgb
image = Image.fromarray(frame)
# image = image.resize((500,500))
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
# shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes[:, :-1]
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
results, score = learner.infer(conf, faces, targets, args.tta)
# print(score)
# print(score[0])
match_score = "{:.2f}".format(score.data[0] * 100)
# print(x)
for idx, bbox in enumerate(bboxes):
if args.score:
frame = draw_box_name(
bbox, names[results[idx] + 1] +
'_{:.2f}'.format(score[idx]), frame)
else:
if float('{:.2f}'.format(score[idx])) > .98:
match_score = None
# name = names[0]
# print(name)
frame = draw_box_name(bbox, "unknown", frame)
else:
name = names[results[idx] + 1]
match_score = match_score
frame = draw_box_name(bbox,
names[results[idx] + 1],
frame)
path = "/home/circle/Downloads/work-care-master/engine/dl/data/facebank" + \
str(name)+"/*.jpg"
filenames = [img for img in glob.glob(path)]
img = cv2.imread(filenames[0])
det_image = cv2.imencode('.jpg', img)[1].tostring()
except:
pass
# print('detect error')
ret, jpeg = cv2.imencode('.jpg', frame)
return jpeg.tostring(), det_image, name, match_score
def fn_face_verify_module():
mtcnn = MTCNN()
print('mtcnn loaded')
learner = face_learner(conf, True)
learner.threshold = args.threshold
if conf.device.type == 'cpu':
learner.load_state(conf, 'cpu_final.pth', True, True)
else:
learner.load_state(conf, 'final.pth', True, True)
learner.model.eval()
print('learner loaded')
if args.update:
targets, names = prepare_facebank(conf,
learner.model,
mtcnn,
tta=args.tta)
print('facebank updated')
else:
targets, names = load_facebank(conf)
print('facebank loaded')
isSuccess, frame = cap.read()
if isSuccess:
try:
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
bboxes = bboxes[:, :
-1] # shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
results, score = learner.infer(conf, faces, targets, args.tta)
for idx, bbox in enumerate(bboxes):
if args.score:
frame = draw_box_name(
bbox,
names[results[idx] + 1] + '_{:.2f}'.format(score[idx]),
frame)
else:
frame = draw_box_name(bbox, names[results[idx] + 1], frame)
except:
print('detect error')
cv2.imshow('face Capture', frame)
if args.save:
video_writer.write(frame)
def authenticuser(path,userid):
conf = get_config(False)
mtcnn = MTCNN()
print('mtcnn loaded')
learner = face_learner(conf, True)
learner.threshold = 1.35
learner.load_state(conf, 'cpu_final.pth', True, True)
learner.model.eval()
print('learner loaded')
targets = load_facebank_user(conf,userid)
names=['Unknown',userid]
print('facebank loaded')
count =0
while True:
frame = cv2.imread(path)
#try:
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit, conf.min_face_size)
bboxes = bboxes[:,:-1] #shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1,-1,1,1] # personal choice
results, score = learner.infer(conf, faces, targets)
for idx,bbox in enumerate(bboxes):
frame = draw_box_name(bbox, names[results[idx] + 1] + '_{:.2f}'.format(100-score[idx]), frame)
result={"_result":"success", "User Verified with":{"confidence": '{:.2f}%'.format(100-score[idx]), "userid": names[results[idx] + 1] , "error": "Success"}}
accuracy.append('{:.2f}'.format(100-score[idx]))
user.append(names[results[idx] + 1])
print( names[results[idx] + 1],'{:.2f}'.format(100-score[idx]))
count=1
#except:
# print('detect error')
if count>0:
break
return result
def infer_general_image(self, image, plot_result=True, tta=False):
# image should be in cv2 format
# If no facebank --> return None
# If plot_result = True --> return annotated image
# If plot_result = False --> return cropped faces and their predicted ID
target_embs = self.targets
names = self.names
if target_embs is None or names is None:
print("No facebank Detected ==> CANT infering!")
return None
origin_image = image
faces, boxes = self.detect_model.detect_face(image)
list_imgs_to_recognise = []
for face in faces:
# align
img, _ = self.alignment_model.align(face)
# Convert to BGR (IMPORTANT)
img = convert_pil_rgb2bgr(img)
list_imgs_to_recognise.append(img)
# recognise
predicted_names, predicted_distances = self.infer(
list_imgs_to_recognise, tta=tta)
if plot_result:
boxes = boxes.astype(int)
boxes = boxes + [-1, -1, 1, 1] # personal choice
for idx, box in enumerate(boxes):
image = draw_box_name(box, predicted_names[idx], origin_image)
return Image.fromarray(image)
return faces, names
def fn_face_verify():
isSuccess, frame = cap.read()
if isSuccess:
try:
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit, conf.min_face_size)
bboxes = bboxes[:, :-1] # shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
results, score = learner.infer(conf, faces, targets, args.tta)
#tolist_results=results.tolist()
#print(tolist_results) #이건 어떻게 바꿀지 잘 모르겠다!
for idx, bbox in enumerate(bboxes):
if args.score:
frame = draw_box_name(bbox, names[results[idx] + 1] + '_{:.2f}'.format(score[idx], frame))
URL=server+"learn"
json_feed={'name':names[results[idx]+1]}
response = requests.post(URL, json=json_feed)
print("----------------")
print(names[results[idx] + 1])
else:
frame = draw_box_name(bbox, names[results[idx] + 1], frame)
URL = server + "learn"
json_feed = {'name': names[results[idx] + 1]}
response = requests.post(URL, json=json_feed)
print(">>>>>>>>>>>>>>>")
print(names[results[idx] + 1])
except:
print('detect error')
cv2.imshow('face Capture',frame)
if args.save:
video_writer.write(frame)
def main(self):
while cap.isOpened():
isSuccess, frame = cap.read()
if isSuccess:
try:
# image = Image.fromarray(frame[...,::-1]) #bgr to rgb
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
bboxes = bboxes[:, :
-1] #shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
results, score = learner.infer(conf, faces, targets,
args.tta)
print(score)
# print(score[0])
for idx, bbox in enumerate(bboxes):
if args.score:
frame = draw_box_name(
bbox, names[results[idx] + 1] +
'_{:.2f}'.format(score[idx]), frame)
else:
if float('{:.2f}'.format(score[idx])) > 1:
name = names[0]
print(name)
frame = draw_box_name(bbox, "unknown", frame)
else:
name = names[results[idx] + 1]
print(name, "extra")
frame = draw_box_name(bbox,
names[results[idx] + 1],
frame)
except:
pass
ret, jpeg = cv2.imencode('.jpg', frame)
return jpeg.tostring()
def get_pic():
isSuccess, frame = cap.read()
if isSuccess:
try:
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
bboxes = bboxes[:, :
-1] # shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
face_list = []
for idx, bbox in enumerate(bboxes):
face_list.append(np.array(faces[idx]).tolist())
URL = server + "register_check"
json_feed_verify = {'face_list': face_list}
start_time = datetime.now()
response = requests.post(URL, json=json_feed_verify)
finish_time = datetime.now()
print('register check time:', finish_time - start_time)
print(response)
check_list = response.json()["check_list"]
for idx, bbox in enumerate(bboxes):
if check_list[idx] == 'unknown':
frame[bbox[1]:bbox[3], bbox[0]:bbox[2]] = cv2.blur(
frame[bbox[1]:bbox[3], bbox[0]:bbox[2]], (23, 23))
else:
frame = draw_box_name(bbox, "known", frame)
cv2.imshow("My Capture", frame)
except:
print("detect error")
if cv2.waitKey(1) & 0xFF == ord('t'):
p = Image.fromarray(frame[..., ::-1])
try:
warped_face = np.array(mtcnn.align(p))[..., ::-1]
re_img = mtcnn.align(p)
tolist_face = np.array(re_img).tolist()
URL = server + "register"
tolist_img = warped_face.tolist()
json_feed = {'face_image': tolist_face}
sregister = datetime.now()
response = requests.post(URL, json=json_feed)
fregister = datetime.now()
except:
print('no face captured')
def get_pic():
isSuccess, frame = cap.read()
if isSuccess:
try:
image = Image.fromarray(frame)
frame_to_server = np.array(image).tolist()
print(np.array(image))
URL = server + "getframe"
json_feed_frame = {'frame_to_server': frame_to_server}
response = requests.post(URL, json=json_feed_frame)
#여기부터
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
bboxes = bboxes[:, :
-1] # shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
face_list = []
for idx, bbox in enumerate(bboxes):
face_list.append(np.array(faces[idx]).tolist())
# start=time.time()
URL = server + "register_check"
json_feed_verify = {'face_list': face_list}
# 보냄.
response = requests.post(URL, json=json_feed_verify)
# 받아오는 것.
print(response)
check_list = response.json()["check_list"]
#여기까지 서버의 getframe측에서 보내주어야함.
for idx, bbox in enumerate(bboxes):
if check_list[idx] == 'unknown':
frame[bbox[1]:bbox[3], bbox[0]:bbox[2]] = cv2.blur(
frame[bbox[1]:bbox[3], bbox[0]:bbox[2]], (23, 23))
else:
frame = draw_box_name(bbox, "known", frame)
cv2.imshow("My Capture", frame)
except:
print("detect error")
# 사진찍은거 넘겨주는 부분,
if cv2.waitKey(1) & 0xFF == ord('t'):
p = Image.fromarray(frame[..., ::-1])
try:
re_img = mtcnn.align(p)
tolist_face = np.array(re_img).tolist()
URL = server + "register"
json_feed = {'face_image': tolist_face}
response = requests.post(URL, json=json_feed)
except:
print('no face captured')
def ddeep():
isSuccess, frame = cap.read()
if isSuccess:
try:
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
bboxes = bboxes[:, :-1]
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1]
face_list = []
for idx, bbox in enumerate(bboxes):
face_list.append(np.array(faces[idx]).tolist())
URL = server + "register_check"
json_feed = {'face_list': face_list}
response = requests.post(URL, json=json_feed)
check_list = response.json()["check_list"]
for idx, bbox in enumerate(bboxes):
if check_list[idx] == 'unknown':
frame[bbox[1]:bbox[3], bbox[0]:bbox[2]] = cv2.blur(
frame[bbox[1]:bbox[3], bbox[0]:bbox[2]], (23, 23))
else:
frame = draw_box_name(bbox, "known", frame)
cv2.imshow('DDeeP', frame)
except:
print("Sorry ")
if cv2.waitKey(1) & 0xFF == ord('r'):
p = Image.fromarray(frame[..., ::-1])
try:
register_face = np.array(mtcnn.align(p))[..., ::-1]
name = 'A'
URL = server + "register"
tolist_face = register_face.tolist()
json_feed = {
'register_image': tolist_face,
'register_name': name
}
response = requests.post(URL, json=json_feed)
print(response)
except:
print('no face captured')
# 키보드에서 c를 누르면 confirm
if cv2.waitKey(0) & 0xFF == ord('c'):
URL = server + "ReadFeature"
params = {'name': 'A'}
res = requests.get(URL, params=params)
res = res.json()
res = res['result']
print(res)
# 키보드에서 n를 누르면 name update
if cv2.waitKey(0) & 0xFF == ord('n'):
URL = server + 'update'
params = {'old_name': 'A', 'new_name': 'NEW'}
res = requests.get(URL, params=params)
print(res.text)
# 키보드에서 u를 누르면 등록된 얼굴을 update할 수 있다.
if cv2.waitKey(0) & 0xFF == ord('u'):
newpic = Image.fromarray(frame[..., ::-1])
new_img = np.array(newpic).tolist()
URL = server + 'update'
json_feed = {'name': 'NEW', 'new_image': new_img}
res = requests.post(URL, json=json_feed)
# 키보드에서 d를 누르면 삭제가능.
if cv2.waitKey(0) & 0xFF == ord('d'):
URL = server + 'delete'
params = {'name': 'NEW'}
res = requests.delete(URL, params=params)
print(res.text)
names.append("{:02d}".format(int(names[-1]) + 1))
for j, bodybbox in enumerate(body_bboxes):
merged_box = merge_box(bodybbox, bboxes[max_inds[j]])
if max_inds[j] != -1:
# match for the previous frame using IOU
iou = np_vec_no_jit_iou(
np.array([bboxes[max_inds[j]]]), bboxes0)
max_iou, max_ind = np.max(iou,
axis=1), np.argmax(iou,
axis=1)
print(iou, max_ind)
if max_iou >= 0.5:
frame = draw_box_name(merged_box,
names0[max_ind[0]], frame)
frame = draw_box_name(bboxes[max_inds[j]],
names0[max_ind[0]], frame)
curr_names.append(names0[max_ind[0]])
else:
frame = draw_box_name(
merged_box, names[results[max_inds[j]] + 1],
frame)
frame = draw_box_name(
bboxes[max_inds[j]],
names[results[max_inds[j]] + 1], frame)
curr_names.append(names[results[max_inds[j]] + 1])
curr_boxes.append(bboxes[max_inds[j]])
if len(curr_names) != 0:
names0 = curr_names
img_size = 112
margin = 0
img_h, img_w, _ = frame.shape
start_time = time.time()
results, score = infer(model=model,
conf=conf,
faces=aligned_faces,
target_embs=targets,
tta=False)
print('Duration: {}'.format(time.time() - start_time))
# results, score = infer(model=model, conf=conf, faces=aligned_faces, target_embs=targets, tta=True)
for idx, bbox in enumerate(bboxes):
x1, y1, x2, y2 = bbox[0], bbox[1], bbox[2], bbox[3]
xw1 = max(int(x1 - margin), 0)
yw1 = max(int(y1 - margin), 0)
xw2 = min(int(x2 + margin), img_w - 1)
yw2 = min(int(y2 + margin), img_h - 1)
bbox = [xw1, yw1, xw2, yw2]
# frame = draw_box_name(bbox, names[results[idx] + 1] + '_{:.2f}'.format(score[idx]), frame)
frame = draw_box_name(bbox, names[results[idx] + 1], frame)
# frame = cv2.resize(frame, dsize=None ,fx=0.25, fy=0.25)
video_writer.write(frame)
# cv2.imshow('window', frame)
# if cv2.waitKey(0) == ord('q'):
# break
cap.release()
video_writer.release()
for fil in path.iterdir():
# if not fil.is_file():
# continue
# else:
print(fil)
frame = cv2.imread(str(fil))
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
bboxes = bboxes[:, :
-1] #shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
results, score = learner.infer(conf, faces, targets, args.tta)
for idx, bbox in enumerate(bboxes):
pred_name = names[results[idx] + 1]
frame = draw_box_name(
bbox, pred_name + '_{:.2f}'.format(score[idx]), frame)
if pred_name in fil.name:
counts[pred_name][1] += 1
else:
orig_name = ''.join([
i for i in fil.name.split('.')[0]
if not i.isdigit()
])
counts[orig_name][0] += 1
# new_name = '_'.join(str(fil).split('/')[-2:])
# print(verify_dir/fil.name)
cv2.imwrite(str(verify_fold_dir / fil.name), frame)
print(counts)
def main(_argv):
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
os.environ['CUDA_VISIBLE_DEVICES'] = FLAGS.gpu
logger = tf.get_logger()
logger.disabled = True
logger.setLevel(logging.FATAL)
set_memory_growth()
cfg = load_yaml(FLAGS.cfg_path)
model = ArcFaceModel(size=cfg['input_size'],
backbone_type=cfg['backbone_type'],
training=False)
ckpt_path = tf.train.latest_checkpoint('./checkpoints/' + cfg['sub_name'])
if ckpt_path is not None:
print("[*] load ckpt from {}".format(ckpt_path))
model.load_weights(ckpt_path)
else:
print("[*] Cannot find ckpt from {}.".format(ckpt_path))
exit()
if FLAGS.update:
print('Face bank updating...')
targets, names = prepare_facebank(cfg, model)
print('Face bank updated')
else:
targets, names = load_facebank(cfg)
print('Face bank loaded')
if FLAGS.video is None:
cap = cv2.VideoCapture(0)
else:
cap = cv2.VideoCapture(str(FLAGS.video))
if FLAGS.save:
video_writer = cv2.VideoWriter('./recording.avi',
cv2.VideoWriter_fourcc(*'XVID'), 10,
(640, 480))
# frame rate 6 due to my laptop is quite slow...
while cap.isOpened():
is_success, frame = cap.read()
if is_success:
img = frame
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
bboxes, landmarks, faces = align_multi(
cfg, img, min_confidence=FLAGS.min_confidence, limits=3)
bboxes = bboxes.astype(int)
embs = []
for face in faces:
if len(face.shape) == 3:
face = np.expand_dims(face, 0)
face = face.astype(np.float32) / 255.
embs.append(l2_norm(model(face)).numpy())
list_min_idx = []
list_score = []
for emb in embs:
dist = [euclidean(emb, target) for target in targets]
min_idx = np.argmin(dist)
list_min_idx.append(min_idx)
list_score.append(dist[int(min_idx)])
list_min_idx = np.array(list_min_idx)
list_score = np.array(list_score)
list_min_idx[list_score > FLAGS.threshold] = -1
for idx, box in enumerate(bboxes):
frame = utils.draw_box_name(box, landmarks[idx],
names[list_min_idx[idx] + 1],
frame)
frame = cv2.resize(frame, (640, 480))
cv2.imshow('face Capture', frame)
key = cv2.waitKey(1) & 0xFF
if FLAGS.save:
video_writer.write(frame)
if key == ord('q'):
break
cap.release()
if FLAGS.save:
video_writer.release()
cv2.destroyAllWindows()
str(conf.data_path / 'recording.mov'),
cv2.VideoWriter_fourcc('m', 'p', '4', 'v'), 6, (1280, 720))
while cap.isOpened():
# start_time = time.time()
_, frame = cap.read()
if frame is None:
break
start_time = time.time()
bbs, fcs = model.find_faces(frame, conf)
for bb, fc in zip(bbs, fcs):
# start_time = time.time()
emb = model.get_feature(fc)
name = get_name(emb, labels)
frame = draw_box_name(bb, name, frame)
cv2.putText(frame, 'FPS: ' + str(1.0 / (time.time() - start_time)),
(50, 50), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 255), 2,
cv2.LINE_AA)
# cv2.imshow('face Capture', frame)
print(name + ' FPS: ' + str(1.0 / (time.time() - start_time)))
# save video
if args.save:
video_writer.write(frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
if args.save:
video_writer.release()
image = Image.open("/media/velab/dati/faces_emore/test/test2.jpeg")
#image.show()
#input("FIRST IMAGE")
# faces rappresenta un arra
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
#faces[0].show()
#input("IMAGE CROPPED")
bboxes = bboxes[:, :
-1] # shape:[10,4],only keep 10 highest possibiity image
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
results, score = learner.infer(conf, faces, targets, args.tta)
print(results, score)
input("RESULT")
image_cv = numpy.array(image)
image_cv = image_cv[:, :, ::-1].copy()
for idx, bbox in enumerate(bboxes):
if args.score:
image_cv = draw_box_name(
bbox, names[results[idx] + 1] + '_{:.2f}'.format(score[idx]),
image_cv)
else:
image_cv = draw_box_name(bbox, names[results[idx] + 1], image_cv)
cv2.imshow('face Capture', image_cv)
cv2.waitKey()
def DDeeP():
isSuccess, frame = cap.read()
if isSuccess:
try:
global name
image = Image.fromarray(frame)
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
bboxes = bboxes[:, :
-1] # shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
face_list = []
for idx, bbox in enumerate(bboxes):
face_list.append(np.array(faces[idx]).tolist())
URL = server + "register_check"
json_feed_verify = {'face_list': face_list}
response = requests.post(URL, json=json_feed_verify)
check_list = response.json()["check_list"]
for idx, bbox in enumerate(bboxes):
if check_list[idx] == 'unknown':
frame[bbox[1]:bbox[3], bbox[0]:bbox[2]] = cv2.blur(
frame[bbox[1]:bbox[3], bbox[0]:bbox[2]], (23, 23))
else:
frame = draw_box_name(bbox, "known", frame)
cv2.imshow("My Capture", frame)
except:
print("detect error")
if cv2.waitKey(1) & 0xFF == ord('t'):
p = Image.fromarray(frame[..., ::-1])
try:
warped_face = np.array(mtcnn.align(p))[..., ::-1]
re_img = mtcnn.align(p)
tolist_face = np.array(re_img).tolist()
#name 이부분에서 입력받도록 해야함.
name = 'Seo Yeon'
URL = server + "register"
tolist_img = warped_face.tolist()
json_feed = {'face_list': tolist_face, 'register_name': name}
response = requests.post(URL, json=json_feed)
except:
print('no face captured')
if cv2.waitKey(0) & 0xFF == ord('c'):
URL = server + "ReadFeature"
params = {'name': name}
res = requests.get(URL, params=params)
res = res.json()
res = res['result']
print(res)
# 키보드에서 n를 누르면 name update 이부분에 대해서는 추후에 업데이트 기능을 만들도록.
if cv2.waitKey(0) & 0xFF == ord('n'):
URL = server + 'update'
new_name = 'NEW'
params = {'old_name': name, 'new_name': new_name}
name = new_name
res = requests.get(URL, params=params)
print(res.text)
# 키보드에서 u를 누르면 등록된 얼굴을 update할 수 있다.
if cv2.waitKey(0) & 0xFF == ord('u'):
newpic = Image.fromarray(frame[..., ::-1])
new_img = np.array(newpic).tolist()
URL = server + 'update'
json_feed = {'name': name, 'new_image': new_img}
res = requests.post(URL, json=json_feed)
# 키보드에서 d를 누르면 삭제가능.
if cv2.waitKey(0) & 0xFF == ord('d'):
URL = server + 'delete'
#이 부분을 변수화해야함.
params = {'name': name}
res = requests.delete(URL, params=params)
print(res.text)
# merge the body box with the matched detection
merged_box = merge_box(bodybbox, detected_box)
# match for the previous frame using IOU
# iou = np_vec_no_jit_iou(np.array([detected_box]), bboxes0)
# max_iou, max_ind = np.max(iou, axis=1), np.argmax(iou, axis=1)
# print(iou, max_ind)
# if max_iou >= 0.5:
# # we have an IoU over 0.5, then there's match. We use the previous frame information
# identity = names0[max_ind[0]]
# frame = draw_box_name(merged_box, identity, frame)
# frame = draw_box_name(detected_box, identity, frame)
# curr_names.append(identity)
# else: # otherwise, look up in the face bank
identity = names[results[matched_detection] + 1]
frame = draw_box_name(merged_box, "", frame)
frame = draw_box_name(detected_box, "", frame)
curr_names.append(identity)
curr_boxes.append(detected_box)
out.append([
int(path.replace(".jpg", "")), identity,
detected_box[0], detected_box[1], detected_box[2],
detected_box[3], merged_box[0], merged_box[1],
merged_box[2], merged_box[3]
])
else:
face_bbox = openpose_face_bboxes[j]
merged_box = merge_box(bodybbox, face_bbox)
frame = draw_box_name(merged_box, "", frame)
out.append([
except:
bboxes = []
faces = []
if len(bboxes) == 0:
print('no face')
continue
else:
bboxes = bboxes[:, :
-1] #shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
results, score = learner.infer(conf, faces, targets, True)
for idx, bbox in enumerate(bboxes):
if args.score:
frame = draw_box_name(
bbox, names[results[idx] + 1] +
'_{:.2f}'.format(score[idx]), frame)
else:
frame = draw_box_name(bbox, names[results[idx] + 1],
frame)
cv2.imshow('img', frame)
cv2.waitKey(10)
video_writer.write(frame)
else:
break
if args.duration != 0:
i += 1
if i % 25 == 0:
print('{} second'.format(i // 25))
if i > 25 * args.duration:
break
shutil.rmtree(bank_path)
os.mkdir(bank_path)
cap = cv2.VideoCapture(0)
model = face_model.FaceModel(conf)
while cap.isOpened():
isSuccess,frame = cap.read()
key = cv2.waitKey(1)&0xFF
bbs, fcs = model.find_faces(frame, conf)
if len(bbs)==1:
frame = draw_box_name(bbs[0], "", frame)
key = cv2.waitKey(1)&0xFF
if key == ord('q') or key == 27:
break
if key == ord('t'):
img = fcs[0][0]
cv2.imwrite(bank_path+'/'+str('{}.jpg'.format(str(datetime.now())[:-7].replace(":","-").replace(" ","-"))), img)
print('da chup')
key = cv2.waitKey(1)&0xFF
if key == ord('q'):
break
cv2.imshow('camera', frame)
cap.release()
cv2.destroyAllWindows()
def test(self,conf,img_dir,update=False,view_score=False,view_error=False):
#Load models
mtcnn = MTCNN()
learner = face_learner(conf, True)
if conf.device.type == 'cpu':
learner.load_state(conf,'cpu_final.pth',True,True)
else:
learner.load_state(conf,'final.pth',True,True)
learner.model.eval()
#Load Facebank
if update:
targets, names = prepare_facebank(conf, learner.model, mtcnn, False)
print('facebank updated')
else:
targets, names = load_facebank(conf)
print('facebank loaded')
#Load Image list
img_list = glob(img_dir + '**/*.jpg')
acc = 0
detect_err=0
fails = []
print(f"{'Found':^15}{'Name':^20}{'Result':^15}{'Score':^15}")
pbar = enumerate(img_list)
pbar = tqdm(pbar, total = len(img_list))
for i, x in pbar:
preds = []
label = str(os.path.dirname(x))
label = os.path.basename(label)
image = Image.open(x)
frame = cv2.imread(x,cv2.IMREAD_COLOR)
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
try:
bboxes, faces = mtcnn.align_multi(image, conf.face_limit, conf.min_face_size)
bboxes = bboxes[:,:-1] #shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1,-1,1,1] # personal choice
results, score = learner.infer(conf, faces, targets, False)
for idx,bbox in enumerate(bboxes):
print(f'{Label}: {score[idx]}')
if view_score:
frame = draw_box_name(bbox, names[results[idx] + 1] + '_{:.2f}'.format(score[idx]), frame)
else:
frame = draw_box_name(bbox, names[results[idx] + 1], frame)
preds.append(names[results[idx]+1])
if label in preds:
acc += 1
else:
fails.append([label,preds])
# Image.fromarray(frame,'RGB').show()
except Exception as ex:
fails.append([label,ex])
detect_err += 1
f = len(bboxes)
tf = str(True if label in preds else False)
t = f'{f:^15}{label:^20}{tf:^15}{acc/(i+1):^15.4}'
pbar.set_description(t)
if detect_err>0:
print(f'Detect Error: {detect_err}')
if view_error:
pp(fails)
else:
print(f'If you want to see details, make veiw_error True.')
print(f'Accuracy: {acc/len(img_list)}')
for i in range(5):
try:
image = Image.fromarray(img[i][..., ::-1]) #bgr to rgb
# image = Image.fromarray(img[i])
print('----------------------------------')
bboxes, faces = mtcnn.align_multi(image, conf.face_limit,
conf.min_face_size)
bboxes = bboxes[:, :
-1] #shape:[10,4],only keep 10 highest possibiity faces
bboxes = bboxes.astype(int)
bboxes = bboxes + [-1, -1, 1, 1] # personal choice
results, score = face_compare(conf, learner.model, faces, targets,
args.tta)
num_face = len(
results) #len(results)가 얼굴개수가나오므로 num_face라는 변수 서연이 만듬.
print(num_face)
for idx, bbox in enumerate(bboxes):
if args.score:
#args.score는 주로 false로 나오기때문에 boundingbox옆에 score가 나오게 하려면 else쪽으로 넣어야함.
img[i] = draw_box_name(bbox, names[results[idx] + 1],
img[i])
else:
img[i] = draw_box_name(
bbox,
names[results[idx] + 1] + '_{:.2f}'.format(score[idx]),
img[i])
except:
print('detect error')
cv2.imwrite('data/output/img_{}.jpg'.format(i), img[i])