def main():
# model files check and download
check_and_download_models(WEIGHT_PATH, MODEL_PATH, REMOTE_PATH)
# net initialize
if args.detector:
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV4,
env_id=args.env_id,
)
else:
detector = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
detector.set_input_shape((1, 3, IMAGE_HEIGHT, IMAGE_WIDTH))
if args.video is not None:
# video mode
recognize_from_video(detector)
else:
# image mode
recognize_from_image(detector)
def main():
# model files check and download
logger.info('Check vehicle-attributes-recognition model...')
check_and_download_models(WEIGHT_PATH, MODEL_PATH, REMOTE_PATH)
if args.video or args.detection:
logger.info('Check object detection model...')
check_and_download_models(DT_WEIGHT_PATH, DT_MODEL_PATH,
DT_REMOTE_PATH)
env_id = args.env_id
# net initialize
net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=env_id)
if args.video or args.detection:
detector = ailia.Detector(
DT_MODEL_PATH,
DT_WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id,
)
else:
detector = None
if args.video:
# video mode
recognize_from_video(net, detector)
else:
# image mode
recognize_from_image(net, detector)
def main():
# model files check and download
logger.info('=== GMM model ===')
check_and_download_models(WEIGHT_GMM_PATH, MODEL_GMM_PATH, REMOTE_PATH)
logger.info('=== TOM model ===')
check_and_download_models(WEIGHT_TOM_PATH, MODEL_TOM_PATH, REMOTE_PATH)
if args.video or not args.keypoints:
logger.info('=== detector model ===')
check_and_download_models(WEIGHT_YOLOV3_PATH, MODEL_YOLOV3_PATH,
REMOTE_YOLOV3_PATH)
logger.info('=== pose model ===')
check_and_download_models(WEIGHT_POSE_PATH, MODEL_POSE_PATH,
REMOTE_POSE_PATH)
if args.video or not args.parse:
logger.info('=== human segmentation model ===')
check_and_download_models(WEIGHT_SEG_PATH, MODEL_SEG_PATH,
REMOTE_SEG_PATH)
# initialize
if args.onnx:
import onnxruntime
GMM_net = onnxruntime.InferenceSession(WEIGHT_GMM_PATH)
TOM_net = onnxruntime.InferenceSession(WEIGHT_TOM_PATH)
else:
GMM_net = ailia.Net(MODEL_GMM_PATH,
WEIGHT_GMM_PATH,
env_id=args.env_id)
TOM_net = ailia.Net(MODEL_TOM_PATH,
WEIGHT_TOM_PATH,
env_id=args.env_id)
if args.video or not args.keypoints:
det_net = ailia.Detector(
MODEL_YOLOV3_PATH,
WEIGHT_YOLOV3_PATH,
80,
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=args.env_id,
)
pose_net = ailia.Net(MODEL_POSE_PATH,
WEIGHT_POSE_PATH,
env_id=args.env_id)
else:
det_net = pose_net = None
if args.video or not args.parse:
seg_net = ailia.Net(MODEL_SEG_PATH,
WEIGHT_SEG_PATH,
env_id=args.env_id)
else:
seg_net = None
if args.video is not None:
# video mode
recognize_from_video(GMM_net, TOM_net, det_net, pose_net, seg_net)
else:
# image mode
recognize_from_image(GMM_net, TOM_net, det_net, pose_net, seg_net)
def recognize_from_image():
# prepare input data
img = load_image(args.input)
print(f'input image shape: {img.shape}')
# net initialize
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_S_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV2,
env_id=env_id
)
detector.set_anchors(ANCHORS)
# compute execution time
for i in range(5):
start = int(round(time.time() * 1000))
detector.compute(img, THRESHOLD, IOU)
end = int(round(time.time() * 1000))
print(f'ailia processing time {end - start} ms')
# plot result
res_img = plot_results(detector, img, COCO_CATEGORY)
cv2.imwrite(args.savepath, res_img)
print('Script finished successfully.')
def recognize_from_image():
# prepare input data
img = load_image(args.input)
print(f'input image shape: {img.shape}')
# net initialize
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
detector = ailia.Detector(MODEL_PATH,
WEIGHT_PATH,
len(FACE_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id)
# inference
print('Start inference...')
if args.benchmark:
print('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
detector.compute(img, THRESHOLD, IOU)
end = int(round(time.time() * 1000))
print(f'\tailia processing time {end - start} ms')
else:
detector.compute(img, THRESHOLD, IOU)
# plot result
res_img = plot_results(detector, img, FACE_CATEGORY)
cv2.imwrite(args.savepath, res_img)
print('Script finished successfully.')
def recognize_from_image():
# prepare input data
img = load_image(args.input)
print(f'input image shape: {img.shape}')
# net initialize
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=args.env_id,
)
if args.detection_width != DETECTION_SIZE or args.detection_height != DETECTION_SIZE:
detector.set_input_shape(args.detection_width, args.detection_height)
# inference
print('Start inference...')
if args.benchmark:
print('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
detector.compute(img, args.threshold, args.iou)
end = int(round(time.time() * 1000))
print(f'\tailia processing time {end - start} ms')
else:
detector.compute(img, args.threshold, args.iou)
# plot result
res_img = plot_results(detector, img, COCO_CATEGORY)
cv2.imwrite(args.savepath, res_img)
print('Script finished successfully.')
def main():
# model files check and download
check_and_download_models(WEIGHT_PATH, MODEL_PATH, REMOTE_PATH)
env_id = args.env_id
if args.detector:
detector = ailia.Detector(MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_BGR,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_INT8,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOX,
env_id=env_id)
if args.detection_width != -1 or args.detection_height != -1:
detector.set_input_shape(args.detection_width,
args.detection_height)
else:
detector = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=env_id)
if args.detection_width != -1 or args.detection_height != -1:
global WIDTH, HEIGHT
WIDTH = args.detection_width
HEIGHT = args.detection_height
detector.set_input_shape((1, 3, HEIGHT, WIDTH))
if args.video is not None:
# video mode
recognize_from_video(detector)
else:
# image mode
recognize_from_image(detector)
def recognize_from_video():
# net initialize
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=args.env_id,
)
if args.detection_width != DETECTION_SIZE or args.detection_height != DETECTION_SIZE:
detector.set_input_shape(
args.detection_width, args.detection_height
)
capture = webcamera_utils.get_capture(args.video)
# create video writer if savepath is specified as video format
if args.savepath != SAVE_IMAGE_PATH:
f_h = int(capture.get(cv2.CAP_PROP_FRAME_HEIGHT))
f_w = int(capture.get(cv2.CAP_PROP_FRAME_WIDTH))
writer = webcamera_utils.get_writer(args.savepath, f_h, f_w)
else:
writer = None
if args.write_prediction:
frame_count = 0
frame_digit = int(math.log10(capture.get(cv2.CAP_PROP_FRAME_COUNT)) + 1)
video_name = os.path.splitext(os.path.basename(args.video))[0]
while(True):
ret, frame = capture.read()
if (cv2.waitKey(1) & 0xFF == ord('q')) or not ret:
break
img = cv2.cvtColor(frame, cv2.COLOR_BGR2BGRA)
detector.compute(img, args.threshold, args.iou)
res_img = plot_results(detector, frame, COCO_CATEGORY, False)
cv2.imshow('frame', res_img)
# save results
if writer is not None:
writer.write(res_img)
# write prediction
if args.write_prediction:
savepath = get_savepath(args.savepath, video_name, post_fix = '_%s' % (str(frame_count).zfill(frame_digit) + '_res'), ext='.png')
pred_file = '%s.txt' % savepath.rsplit('.', 1)[0]
write_predictions(pred_file, detector, frame, COCO_CATEGORY)
frame_count += 1
capture.release()
cv2.destroyAllWindows()
if writer is not None:
writer.release()
logger.info('Script finished successfully.')
def recognize_from_image():
# net initialize
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=args.env_id,
)
if args.detection_width != DETECTION_SIZE or args.detection_height != DETECTION_SIZE:
detector.set_input_shape(
args.detection_width, args.detection_height
)
if args.profile:
detector.set_profile_mode(True)
# input image loop
for image_path in args.input:
# prepare input data
logger.info(image_path)
img = load_image(image_path)
logger.debug(f'input image shape: {img.shape}')
# inference
logger.info('Start inference...')
if args.benchmark:
logger.info('BENCHMARK mode')
total_time = 0
for i in range(args.benchmark_count):
start = int(round(time.time() * 1000))
detector.compute(img, args.threshold, args.iou)
end = int(round(time.time() * 1000))
if i != 0:
total_time = total_time + (end - start)
logger.info(f'\tailia processing time {end - start} ms')
logger.info(f'\taverage time {total_time / (args.benchmark_count-1)} ms')
else:
detector.compute(img, args.threshold, args.iou)
# plot result
res_img = plot_results(detector, img, COCO_CATEGORY)
savepath = get_savepath(args.savepath, image_path)
logger.info(f'saved at : {savepath}')
cv2.imwrite(savepath, res_img)
# write prediction
if args.write_prediction:
pred_file = '%s.txt' % savepath.rsplit('.', 1)[0]
write_predictions(pred_file, detector, img, COCO_CATEGORY)
if args.profile:
print(detector.get_summary())
logger.info('Script finished successfully.')
def recognize_from_video():
# net initialize
if args.detector == True:
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_S_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV2,
env_id=args.env_id,
)
detector.set_anchors(ANCHORS)
else:
net = ailia.Net(None, WEIGHT_PATH)
capture = webcamera_utils.get_capture(args.video)
# create video writer if savepath is specified as video format
if args.savepath != SAVE_IMAGE_PATH:
f_h = int(capture.get(cv2.CAP_PROP_FRAME_HEIGHT))
f_w = int(capture.get(cv2.CAP_PROP_FRAME_WIDTH))
writer = webcamera_utils.get_writer(args.savepath, f_h, f_w)
else:
writer = None
while (True):
ret, frame = capture.read()
if (cv2.waitKey(1) & 0xFF == ord('q')) or not ret:
break
if args.detector:
img = cv2.cvtColor(frame, cv2.COLOR_BGR2BGRA)
detector.compute(img, THRESHOLD, IOU)
res_img = plot_results(detector, frame, CATEGORY, False)
else:
img_PIL = Image.fromarray(frame)
img = cv2.cvtColor(frame, cv2.COLOR_RGB2BGR)
img = cv2.resize(img, (416, 416))
img = img.transpose((2, 0, 1)) / 255
img = img[np.newaxis, :, :, :].astype(np.float32)
results = net.run([img])
results = torch.FloatTensor(results[0])
output_img = detect(img_PIL, results, video=True)
res_img = np.array(output_img, dtype=np.uint8)
cv2.imshow('frame', res_img)
# save results
if writer is not None:
writer.write(res_img)
capture.release()
cv2.destroyAllWindows()
if writer is not None:
writer.release()
logger.info('Script finished successfully.')
def recognize_from_video():
# net initialize
detector = ailia.Detector(MODEL_PATH,
WEIGHT_PATH,
len(FACE_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=RANGE,
algorithm=ALGORITHM,
env_id=args.env_id)
capture = webcamera_utils.get_capture(args.video)
if args.savepath != SAVE_IMAGE_PATH:
writer = webcamera_utils.get_writer(
args.savepath,
IMAGE_HEIGHT,
IMAGE_WIDTH,
fps=capture.get(cv2.CAP_PROP_FPS),
)
else:
writer = None
while (True):
ret, frame = capture.read()
if (cv2.waitKey(1) & 0xFF == ord('q')) or not ret:
break
_, resized_img = webcamera_utils.adjust_frame_size(
frame, IMAGE_HEIGHT, IMAGE_WIDTH)
img = cv2.cvtColor(resized_img, cv2.COLOR_BGR2BGRA)
detector.compute(img, THRESHOLD, IOU)
detections = []
for idx in range(detector.get_object_count()):
obj = detector.get_object(idx)
detections.append(obj)
detections = nms_between_categories(detections,
frame.shape[1],
frame.shape[0],
categories=[0, 1],
iou_threshold=IOU)
res_img = plot_results(detections, resized_img, FACE_CATEGORY, False)
cv2.imshow('frame', res_img)
# save results
if writer is not None:
writer.write(res_img)
capture.release()
cv2.destroyAllWindows()
if writer is not None:
writer.release()
logger.info('Script finished successfully.')
def init_detector(env_id):
detector = ailia.Detector(DT_MODEL_PATH,
DT_WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id)
return detector
def recognize_from_image():
# net initialize
categories = 80
threshold = 0.4
iou = 0.45
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
categories,
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_SSD,
env_id=args.env_id,
)
if args.profile:
detector.set_profile_mode(True)
# input image loop
for image_path in args.input:
# prepare input data
logger.info(image_path)
org_img = load_image(
image_path,
(IMAGE_HEIGHT, IMAGE_WIDTH),
normalize_type='None',
)
if org_img.shape[2] == 3:
org_img = cv2.cvtColor(org_img, cv2.COLOR_RGB2BGRA)
# inference
logger.info('Start inference...')
if args.benchmark:
logger.info('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
detector.compute(org_img, threshold, iou)
end = int(round(time.time() * 1000))
logger.info(f'\tailia processing time {end - start} ms')
else:
detector.compute(org_img, threshold, iou)
# postprocessing
res_img = plot_results(detector, org_img, VOC_CATEGORY)
savepath = get_savepath(args.savepath, image_path)
logger.info(f'saved at : {savepath}')
cv2.imwrite(savepath, res_img)
if args.profile:
print(detector.get_summary())
logger.info('Script finished successfully.')
def main():
# model files check and download
check_and_download_models(WEIGHT_PATH, MODEL_PATH, REMOTE_PATH)
# check folder existing
if not os.path.exists(args.input):
print("error : directory not found "+args.input)
sys.exit(1)
if not os.path.exists(args.output):
os.mkdir(args.output)
# net initialize
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
if args.arch == 'blazeface':
detector = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=env_id)
else:
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(FACE_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id
)
# process images
no = 0
for src_dir, dirs, files in os.walk(args.input):
files = sorted(files)
for file_ in files:
root, ext = os.path.splitext(file_)
if file_==".DS_Store":
continue
if file_=="Thumbs.db":
continue
if not(ext == ".jpg" or ext == ".png" or ext == ".bmp"):
continue
print(src_dir+"/"+file_)
folders=src_dir.split("/")
folder=folders[len(folders)-1]
dst_dir = args.output+"/"+folder
if not os.path.exists(dst_dir):
os.mkdir(dst_dir)
dst_path= dst_dir+ "/"+str(no)+".jpg"
recognize_from_image(detector,dst_path,src_dir,file_)
no=no+1
def recognize_from_image():
# net initialize
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=args.env_id,
)
pose = ailia.Net(POSE_MODEL_PATH, POSE_WEIGHT_PATH, env_id=args.env_id)
# input image loop
for image_path in args.input:
# prepare input data
logger.info(image_path)
img = load_image(image_path)
logger.debug(f'input image shape: {img.shape}')
# inference
logger.info('Start inference...')
detector.compute(img, THRESHOLD, IOU)
# pose estimation
if args.benchmark:
logger.info('BENCHMARK mode')
total_time = 0
for i in range(args.benchmark_count):
start = int(round(time.time() * 1000))
pose_detections = pose_estimation(detector, pose, img)
end = int(round(time.time() * 1000))
logger.info(
f'\tailia processing detection time {end - start} ms')
if i != 0:
total_time = total_time + (end - start)
logger.info(
f'\taverage detection time {total_time / (args.benchmark_count-1)} ms'
)
else:
pose_detections = pose_estimation(detector, pose, img)
# plot result
res_img = plot_results(detector, pose, img, COCO_CATEGORY,
pose_detections)
savepath = get_savepath(args.savepath, image_path)
logger.info(f'saved at : {savepath}')
cv2.imwrite(savepath, res_img)
logger.info('Script finished successfully.')
def recognize_from_video():
# net initialize
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(FACE_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=RANGE,
algorithm=ALGORITHM,
env_id=env_id
)
if args.video == '0':
print('[INFO] Webcam mode is activated')
capture = cv2.VideoCapture(0)
if not capture.isOpened():
print("[ERROR] webcamera not found")
sys.exit(1)
else:
if check_file_existance(args.video):
capture = cv2.VideoCapture(args.video)
while(True):
ret, frame = capture.read()
if cv2.waitKey(1) & 0xFF == ord('q'):
break
if not ret:
continue
_, resized_img = adjust_frame_size(frame, IMAGE_HEIGHT, IMAGE_WIDTH)
img = cv2.cvtColor(resized_img, cv2.COLOR_BGR2BGRA)
detector.compute(img, THRESHOLD, IOU)
detections = []
for idx in range(detector.get_object_count()):
obj = detector.get_object(idx)
detections.append(obj)
detections=nms_between_categories(detections,frame.shape[1],frame.shape[0],categories=[0,1],iou_threshold=IOU)
res_img = plot_results(detections, resized_img, FACE_CATEGORY, False)
cv2.imshow('frame', res_img)
capture.release()
cv2.destroyAllWindows()
print('Script finished successfully.')
def recognize_from_image():
# net initialize
detector = ailia.Detector(MODEL_PATH,
WEIGHT_PATH,
len(FACE_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=RANGE,
algorithm=ALGORITHM,
env_id=args.env_id)
# input image loop
for image_path in args.input:
# prepare input data
logger.info(image_path)
img = load_image(image_path)
logger.debug(f'input image shape: {img.shape}')
# inference
logger.info('Start inference...')
if args.benchmark:
logger.info('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
detector.compute(img, THRESHOLD, IOU)
end = int(round(time.time() * 1000))
logger.info(f'\tailia processing time {end - start} ms')
else:
detector.compute(img, THRESHOLD, IOU)
# nms
detections = []
for idx in range(detector.get_object_count()):
obj = detector.get_object(idx)
detections.append(obj)
detections = nms_between_categories(
detections,
img.shape[1],
img.shape[0],
categories=[0, 1],
iou_threshold=IOU,
)
# plot result
res_img = plot_results(detections, img, FACE_CATEGORY)
savepath = get_savepath(args.savepath, image_path)
logger.info(f'saved at : {savepath}')
cv2.imwrite(savepath, res_img)
logger.info('Script finished successfully.')
def __init__(self):
check_and_download_models(WEIGHT_PATH_YOLO, MODEL_PATH_YOLO,
REMOTE_PATH_YOLO)
# net initialize
env_id = ailia.get_gpu_environment_id()
self.detector = detector = ailia.Detector(
MODEL_PATH_YOLO,
WEIGHT_PATH_YOLO,
len(FACE_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id)
def recognize_from_video():
# net initialize
categories = 80
threshold = 0.4
iou = 0.45
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
categories,
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_SSD,
env_id=args.env_id,
)
capture = webcamera_utils.get_capture(args.video)
# create video writer if savepath is specified as video format
if args.savepath != SAVE_IMAGE_PATH:
writer = webcamera_utils.get_writer(
args.savepath, IMAGE_HEIGHT, IMAGE_WIDTH
)
else:
writer = None
while(True):
ret, frame = capture.read()
if (cv2.waitKey(1) & 0xFF == ord('q')) or not ret:
break
_, resized_img = webcamera_utils.adjust_frame_size(
frame, IMAGE_HEIGHT, IMAGE_WIDTH
)
img = cv2.cvtColor(resized_img, cv2.COLOR_BGR2BGRA)
detector.compute(img, threshold, iou)
res_img = plot_results(detector, resized_img, VOC_CATEGORY, False)
cv2.imshow('frame', res_img)
# save results
if writer is not None:
writer.write(res_img)
capture.release()
cv2.destroyAllWindows()
if writer is not None:
writer.release()
logger.info('Script finished successfully.')
def recognize_from_video():
# net initialize
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=args.env_id,
)
if args.detection_width != DETECTION_SIZE or args.detection_height != DETECTION_SIZE:
detector.set_input_shape(
args.detection_width, args.detection_height
)
capture = webcamera_utils.get_capture(args.video)
# create video writer if savepath is specified as video format
if args.savepath != SAVE_IMAGE_PATH:
f_h = int(capture.get(cv2.CAP_PROP_FRAME_HEIGHT))
f_w = int(capture.get(cv2.CAP_PROP_FRAME_WIDTH))
writer = webcamera_utils.get_writer(args.savepath, f_h, f_w)
else:
writer = None
while(True):
ret, frame = capture.read()
if (cv2.waitKey(1) & 0xFF == ord('q')) or not ret:
break
img = cv2.cvtColor(frame, cv2.COLOR_BGR2BGRA)
detector.compute(img, args.threshold, args.iou)
res_img = plot_results(detector, frame, COCO_CATEGORY, False)
cv2.imshow('frame', res_img)
# save results
if writer is not None:
writer.write(res_img)
capture.release()
cv2.destroyAllWindows()
if writer is not None:
writer.release()
logger.info('Script finished successfully.')
def recognize_from_image():
# net initialize
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_S_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV2,
env_id=args.env_id,
)
detector.set_anchors(ANCHORS)
if args.profile:
detector.set_profile_mode(True)
# input image loop
for image_path in args.input:
# prepare input data
logger.info(image_path)
img = load_image(image_path)
logger.debug(f'input image shape: {img.shape}')
# inference
logger.info('Start inference...')
if args.benchmark:
logger.info('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
detector.compute(img, THRESHOLD, IOU)
end = int(round(time.time() * 1000))
logger.info(f'\tailia processing time {end - start} ms')
else:
detector.compute(img, THRESHOLD, IOU)
# plot result
res_img = plot_results(detector, img, COCO_CATEGORY)
savepath = get_savepath(args.savepath, image_path)
logger.info(f'saved at : {savepath}')
cv2.imwrite(savepath, res_img)
if args.profile:
print(detector.get_summary())
logger.info('Script finished successfully.')
def recognize_from_video():
# net initialize
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
detector = ailia.Detector(
MODEL_PATH,
WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_S_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV2,
env_id=env_id
)
detector.set_anchors(ANCHORS)
if args.video == '0':
print('[INFO] Webcam mode is activated')
capture = cv2.VideoCapture(0)
if not capture.isOpened():
print("[ERROR] webcamera not found")
sys.exit(1)
else:
if check_file_existance(args.video):
capture = cv2.VideoCapture(args.video)
while(True):
ret, frame = capture.read()
if cv2.waitKey(1) & 0xFF == ord('q'):
break
if not ret:
continue
_, resized_img = adjust_frame_size(frame, IMAGE_HEIGHT, IMAGE_WIDTH)
img = cv2.cvtColor(resized_img, cv2.COLOR_RGB2BGRA)
detector.compute(img, THRESHOLD, IOU)
res_img = plot_results(detector, resized_img, COCO_CATEGORY, False)
cv2.imshow('frame', res_img)
capture.release()
cv2.destroyAllWindows()
print('Script finished successfully.')
def main():
# model files check and download
detector = True
if detector:
logger.info('=== detector model ===')
check_and_download_models(WEIGHT_YOLOV3_PATH, MODEL_YOLOV3_PATH,
REMOTE_YOLOV3_PATH)
logger.info('=== animalpose model ===')
info = {
'hrnet32': (WEIGHT_HRNET_W32_PATH, MODEL_HRNET_W32_PATH),
'hrnet48': (WEIGHT_HRNET_W48_PATH, MODEL_HRNET_W48_PATH),
'res50': (WEIGHT_RESNET_50_PATH, MODEL_RESNET_50_PATH),
'res101': (WEIGHT_RESNET_101_PATH, MODEL_RESNET_101_PATH),
'res152': (WEIGHT_RESNET_152_PATH, MODEL_RESNET_152_PATH),
}
weight_path, model_path = info[args.model]
check_and_download_models(weight_path, model_path, REMOTE_PATH)
env_id = args.env_id
# initialize
if detector:
det_net = ailia.Detector(
MODEL_YOLOV3_PATH,
WEIGHT_YOLOV3_PATH,
80,
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id,
)
else:
det_net = None
net = ailia.Net(model_path, weight_path, env_id=env_id)
if args.video is not None:
# video mode
recognize_from_video(net, det_net)
else:
# image mode
recognize_from_image(net, det_net)
def recognize_from_image():
# prepare input data
org_img = load_image(
args.input,
(IMAGE_HEIGHT, IMAGE_WIDTH),
normalize_type='None',
)
if org_img.shape[2] == 3:
org_img = cv2.cvtColor(org_img, cv2.COLOR_BGR2BGRA)
# net initialize
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
categories = 80
threshold = 0.4
iou = 0.45
detector = ailia.Detector(MODEL_PATH,
WEIGHT_PATH,
categories,
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_SSD,
env_id=env_id)
# inference
print('Start inference...')
if args.benchmark:
print('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
detector.compute(org_img, threshold, iou)
end = int(round(time.time() * 1000))
print(f'\tailia processing time {end - start} ms')
else:
detector.compute(org_img, threshold, iou)
# postprocessing
res_img = plot_results(detector, org_img, VOC_CATEGORY)
cv2.imwrite(args.savepath, res_img)
print('Script finished successfully.')
def main():
# model files check and download
logger.info("=== YOLOv3 model ===")
check_and_download_models(WEIGHT_YOLOV3_PATH, MODEL_YOLOV3_PATH,
REMOTE_YOLOV3_PATH)
logger.info("=== HRNet model ===")
check_and_download_models(WEIGHT_POSE_PATH, MODEL_POSE_PATH, REMOTE_PATH)
logger.info("=== GAST model ===")
check_and_download_models(WEIGHT_27FRAME_17JOINT_PATH,
MODEL_27FRAME_17JOINT_PATH, REMOTE_PATH)
num_person = args.num_person
# net initialize
detector = ailia.Detector(
MODEL_YOLOV3_PATH,
WEIGHT_YOLOV3_PATH,
80,
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=args.env_id,
)
pose_net = ailia.Net(MODEL_POSE_PATH, WEIGHT_POSE_PATH, env_id=args.env_id)
if not args.onnx:
net = ailia.Net(MODEL_27FRAME_17JOINT_PATH,
WEIGHT_27FRAME_17JOINT_PATH,
env_id=args.env_id)
else:
import onnxruntime
net = onnxruntime.InferenceSession(WEIGHT_27FRAME_17JOINT_PATH)
info = {
"yolo_model": detector,
"pose_model": pose_net,
"num_person": num_person,
}
recognize_from_video(net, info)
def compare_video():
# prepare base image
tracks = []
# net initialize
net = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
# detector initialize
if args.face == "blazeface":
detector = ailia.Net(FACE_MODEL_PATH,
FACE_WEIGHT_PATH,
env_id=args.env_id)
else:
detector = ailia.Detector(FACE_MODEL_PATH,
FACE_WEIGHT_PATH,
len(FACE_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=FACE_RANGE,
algorithm=FACE_ALGORITHM,
env_id=args.env_id)
# web camera
capture = webcamera_utils.get_capture(args.video)
# ui buffer
ui_width = capture.get(
cv2.CAP_PROP_FRAME_WIDTH) + IMAGE_WIDTH / 4 * FACE_TRACK_T
ui_height = max(capture.get(cv2.CAP_PROP_FRAME_HEIGHT),
IMAGE_HEIGHT / 4 * 2 * 8)
ui = np.zeros((int(ui_height), int(ui_width), 3), np.uint8)
frame_no = 0
# writer
writer = None
if args.savepath is not None:
writer = webcamera_utils.get_writer(
args.savepath,
ui.shape[0],
ui.shape[1],
fps=capture.get(cv2.CAP_PROP_FPS),
)
# inference loop
while (True):
ret, frame = capture.read()
if (cv2.waitKey(1) & 0xFF == ord('q')) or not ret:
break
# get frame size
h, w = frame.shape[0], frame.shape[1]
# get faces from image
detections = get_faces(detector, frame, w, h)
# track face
face_identification(tracks, net, detections, frame_no)
frame_no = frame_no + 1
# display result
ui[:, :, :] = 0
ui[0:h, 0:w, :] = frame[:, :, :]
display_detections(ui, w, h, detections)
display_tracks(ui, w, h, tracks)
# show
cv2.imshow('arcface', ui)
if writer is not None:
writer.write(ui)
if writer is not None:
writer.release()
capture.release()
cv2.destroyAllWindows()
logger.info('Script finished successfully.')
def recognize_from_video():
try:
print('[INFO] Webcam mode is activated')
RECORD_TIME = 80
capture = cv2.VideoCapture(int(args.video))
if not capture.isOpened():
print("[ERROR] webcamera not found")
sys.exit(1)
except ValueError:
if check_file_existance(args.video):
capture = cv2.VideoCapture(args.video)
frame_rate = capture.get(cv2.CAP_PROP_FPS)
if FRAME_SKIP:
action_recognize_fps = int(args.fps)
else:
action_recognize_fps = frame_rate
if args.savepath != "":
size = (int(capture.get(cv2.CAP_PROP_FRAME_WIDTH)),
int(capture.get(cv2.CAP_PROP_FRAME_HEIGHT)))
fmt = cv2.VideoWriter_fourcc('m', 'p', '4', 'v')
writer = cv2.VideoWriter(args.savepath, fmt, action_recognize_fps,
size)
else:
writer = None
# pose estimation
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
if args.arch == "lw_human_pose":
pose = ailia.PoseEstimator(MODEL_PATH,
WEIGHT_PATH,
env_id=env_id,
algorithm=ALGORITHM)
detector = None
else:
detector = ailia.Detector(DETECTOR_MODEL_PATH,
DETECTOR_WEIGHT_PATH,
len(COCO_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id)
pose = ailia.Net(POSE_MODEL_PATH, POSE_WEIGHT_PATH, env_id=env_id)
# tracker class instance
extractor = ailia.Net(EX_MODEL_PATH, EX_WEIGHT_PATH, env_id=env_id)
metric = NearestNeighborDistanceMetric("cosine", MAX_COSINE_DISTANCE,
NN_BUDGET)
tracker = Tracker(metric, max_iou_distance=0.7, max_age=70, n_init=3)
# action recognition
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
model = ailia.Net(ACTION_MODEL_PATH, ACTION_WEIGHT_PATH, env_id=env_id)
action_data = {}
frame_nb = int(capture.get(cv2.CAP_PROP_FRAME_COUNT))
idx_frame = 0
time_start = time.time()
while (True):
time_curr = time.time()
if args.video == '0' and time_curr - time_start > RECORD_TIME:
break
ret, frame = capture.read()
if cv2.waitKey(1) & 0xFF == ord('q'):
break
if (not ret) or (frame_nb >= 1 and idx_frame >= frame_nb):
break
if FRAME_SKIP:
mod = round(frame_rate / action_recognize_fps)
if mod >= 1:
if idx_frame % mod != 0:
idx_frame = idx_frame + 1
continue
input_image, input_data = adjust_frame_size(
frame,
frame.shape[0],
frame.shape[1],
)
input_data = cv2.cvtColor(input_data, cv2.COLOR_BGR2BGRA)
# inferece
if args.arch == "lw_human_pose":
_ = pose.compute(input_data)
else:
detector.compute(input_data, THRESHOLD, IOU)
# deepsort format
h, w = input_image.shape[0], input_image.shape[1]
if args.arch == "lw_human_pose":
bbox_xywh, cls_conf, cls_ids = get_detector_result_lw_human_pose(
pose, h, w)
else:
bbox_xywh, cls_conf, cls_ids = get_detector_result(detector, h, w)
mask = cls_ids == 0
bbox_xywh = bbox_xywh[mask]
# bbox dilation just in case bbox too small,
# delete this line if using a better pedestrian detector
if args.arch == "pose_resnet":
# bbox_xywh[:, 3:] *= 1.2 #May need to be removed in the future
cls_conf = cls_conf[mask]
# do tracking
img_crops = []
for box in bbox_xywh:
x1, y1, x2, y2 = xywh_to_xyxy(box, h, w)
img_crops.append(input_image[y1:y2, x1:x2])
if img_crops:
# preprocess
img_batch = np.concatenate([
normalize_image(resize(img), 'ImageNet')[np.newaxis, :, :, :]
for img in img_crops
],
axis=0).transpose(0, 3, 1, 2)
# TODO better to pass a batch at once
# features = extractor.predict(img_batch)
features = []
for img in img_batch:
features.append(extractor.predict(img[np.newaxis, :, :, :])[0])
features = np.array(features)
else:
features = np.array([])
bbox_tlwh = xywh_to_tlwh(bbox_xywh)
detections = [
Detection(bbox_tlwh[i], conf, features[i])
for i, conf in enumerate(cls_conf) if conf > MIN_CONFIDENCE
]
# run on non-maximum supression
boxes = np.array([d.tlwh for d in detections])
scores = np.array([d.confidence for d in detections])
nms_max_overlap = 1.0
indices = non_max_suppression(boxes, nms_max_overlap, scores)
detections = [detections[i] for i in indices]
# update tracker
tracker.predict()
tracker.update(detections)
# update bbox identities
outputs = []
for track in tracker.tracks:
if not track.is_confirmed() or track.time_since_update > 1:
continue
box = track.to_tlwh()
x1, y1, x2, y2 = tlwh_to_xyxy(box, h, w)
track_id = track.track_id
outputs.append(np.array([x1, y1, x2, y2, track_id], dtype=np.int))
if len(outputs) > 0:
outputs = np.stack(outputs, axis=0)
# action detection
actions = []
persons = []
if len(outputs) > 0:
bbox_xyxy = outputs[:, :4]
identities = outputs[:, -1]
for i, box in enumerate(bbox_xyxy):
id = identities[i]
if not (id in action_data):
action_data[id] = np.zeros(
(ailia.POSE_KEYPOINT_CNT - 1, TIME_RANGE, 3))
# action recognition
action, person = action_recognition(box, input_image, pose,
detector, model,
action_data[id])
actions.append(action)
persons.append(person)
# draw box for visualization
if len(outputs) > 0:
bbox_tlwh = []
bbox_xyxy = outputs[:, :4]
identities = outputs[:, -1]
frame = draw_boxes(input_image, bbox_xyxy, identities, actions,
action_data, (0, 0))
for bb_xyxy in bbox_xyxy:
bbox_tlwh.append(xyxy_to_tlwh(bb_xyxy))
# draw skelton
for person in persons:
if person != None:
display_result(input_image, person)
if writer is not None:
writer.write(input_image)
# show progress
if idx_frame == "0":
print()
print("\r" + str(idx_frame + 1) + " / " + str(frame_nb), end="")
if idx_frame == frame_nb - 1:
print()
cv2.imshow('frame', input_image)
idx_frame = idx_frame + 1
if writer is not None:
writer.release()
capture.release()
cv2.destroyAllWindows()
print('Script finished successfully.')
def recognize_from_video():
# net initialize
env_id = ailia.get_gpu_environment_id()
if args.env_id is not None:
count = ailia.get_environment_count()
if count > args.env_id:
env_id = args.env_id
else:
print(f'specified env_id: {args.env_id} cannot found error')
print(f'env_id: {env_id}')
detector = ailia.Detector(MODEL_PATH,
WEIGHT_PATH,
len(HAND_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id)
hand = ailia.PoseEstimator(HAND_MODEL_PATH,
HAND_WEIGHT_PATH,
env_id=env_id,
algorithm=HAND_ALGORITHM)
hand.set_threshold(0.1)
ailia_input_w = detector.get_input_shape()[3]
ailia_input_h = detector.get_input_shape()[2]
capture = get_capture(args.video)
# create video writer if savepath is specified as video format
if args.savepath != SAVE_PATH:
f_h = int(capture.get(cv2.CAP_PROP_FRAME_HEIGHT))
f_w = int(capture.get(cv2.CAP_PROP_FRAME_WIDTH))
save_h, save_w = calc_adjust_fsize(f_h, f_w, ailia_input_h,
ailia_input_w)
writer = get_writer(args.savepath, save_h, save_w)
else:
writer = None
while (True):
ret, frame = capture.read()
if (cv2.waitKey(1) & 0xFF == ord('q')) or not ret:
break
img = cv2.cvtColor(frame, cv2.COLOR_BGR2BGRA)
detector.compute(img, THRESHOLD, IOU)
h, w = img.shape[0], img.shape[1]
count = detector.get_object_count()
for idx in range(count):
# get detected hand
obj = detector.get_object(idx)
margin = 1.0
cx = (obj.x + obj.w / 2) * w
cy = (obj.y + obj.h / 2) * h
cw = max(obj.w * w, obj.h * h) * margin
fx = max(cx - cw / 2, 0)
fy = max(cy - cw / 2, 0)
fw = min(cw, w - fx)
fh = min(cw, h - fy)
top_left = (int(fx), int(fy))
bottom_right = (int(fx + fw), int(fy + fh))
# display detected hand
color = hsv_to_rgb(0, 255, 255)
cv2.rectangle(frame, top_left, bottom_right, color, 4)
# get detected face
crop_img = img[top_left[1]:bottom_right[1],
top_left[0]:bottom_right[0], 0:4]
if crop_img.shape[0] <= 0 or crop_img.shape[1] <= 0:
continue
# inference
_ = hand.compute(crop_img.astype(np.uint8, order='C'))
# postprocessing
display_result(frame, hand, top_left, bottom_right)
cv2.imshow('frame', frame)
# save results
if writer is not None:
writer.write(frame)
capture.release()
cv2.destroyAllWindows()
if writer is not None:
writer.release()
print('Script finished successfully.')
def recognize_from_video():
# net initialize
env_id = ailia.get_gpu_environment_id()
print(f'env_id: {env_id}')
detector = ailia.Detector(MODEL_PATH,
WEIGHT_PATH,
len(HAND_CATEGORY),
format=ailia.NETWORK_IMAGE_FORMAT_RGB,
channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST,
range=ailia.NETWORK_IMAGE_RANGE_U_FP32,
algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3,
env_id=env_id)
hand = ailia.PoseEstimator(HAND_MODEL_PATH,
HAND_WEIGHT_PATH,
env_id=env_id,
algorithm=HAND_ALGORITHM)
hand.set_threshold(0.1)
if args.video == '0':
print('[INFO] Webcam mode is activated')
capture = cv2.VideoCapture(0)
if not capture.isOpened():
print("[ERROR] webcamera not found")
sys.exit(1)
else:
if check_file_existance(args.video):
capture = cv2.VideoCapture(args.video)
while (True):
ret, frame = capture.read()
if cv2.waitKey(1) & 0xFF == ord('q'):
break
if not ret:
continue
img = cv2.cvtColor(frame, cv2.COLOR_BGR2BGRA)
detector.compute(img, THRESHOLD, IOU)
h, w = img.shape[0], img.shape[1]
count = detector.get_object_count()
for idx in range(count):
# get detected hand
obj = detector.get_object(idx)
margin = 1.0
cx = (obj.x + obj.w / 2) * w
cy = (obj.y + obj.h / 2) * h
cw = max(obj.w * w, obj.h * h) * margin
fx = max(cx - cw / 2, 0)
fy = max(cy - cw / 2, 0)
fw = min(cw, w - fx)
fh = min(cw, h - fy)
top_left = (int(fx), int(fy))
bottom_right = (int(fx + fw), int(fy + fh))
# display detected hand
color = hsv_to_rgb(0, 255, 255)
cv2.rectangle(frame, top_left, bottom_right, color, 4)
# get detected face
crop_img = img[top_left[1]:bottom_right[1],
top_left[0]:bottom_right[0], 0:4]
if crop_img.shape[0] <= 0 or crop_img.shape[1] <= 0:
continue
# inferece
_ = hand.compute(crop_img.astype(np.uint8, order='C'))
# postprocessing
display_result(frame, hand, top_left, bottom_right)
cv2.imshow('frame', frame)
capture.release()
cv2.destroyAllWindows()
print('Script finished successfully.')
model_path = sys.argv[1]+".prototxt"
weight_path = sys.argv[1]
classes_path = sys.argv[2]
img_path = sys.argv[3]
with open(classes_path) as f:
class_names = f.readlines()
class_names = [c.strip() for c in class_names]
print("loading ...");
# detector initialize
env_id = ailia.get_gpu_environment_id()
categories = len(class_names)
detector = ailia.Detector(model_path, weight_path, categories, format=ailia.NETWORK_IMAGE_FORMAT_RGB, channel=ailia.NETWORK_IMAGE_CHANNEL_FIRST, range=ailia.NETWORK_IMAGE_RANGE_U_FP32, algorithm=ailia.DETECTOR_ALGORITHM_YOLOV3, env_id=env_id)
# load input image and convert to BGRA
img = cv2.imread( img_path, cv2.IMREAD_UNCHANGED )
if img.shape[2] == 3 :
img = cv2.cvtColor( img, cv2.COLOR_BGR2BGRA )
elif img.shape[2] == 1 :
img = cv2.cvtColor( img, cv2.COLOR_GRAY2BGRA )
print( "img.shape=" + str(img.shape) )
work = img
w = img.shape[1]
h = img.shape[0]
print("inferencing ...");