def recognize_from_image(detector):
# prepare input data
org_img = load_image(args.input)
print(f'input image shape: {org_img.shape}')
org_img = cv2.cvtColor(org_img, cv2.COLOR_BGRA2BGR)
img = letterbox_convert(org_img, (IMAGE_HEIGHT, IMAGE_WIDTH))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img = np.transpose(img, [2, 0, 1])
img = img.astype(np.float32) / 255
img = np.expand_dims(img, 0)
# inference
print('Start inference...')
if args.benchmark:
print('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
output = detector.predict([img])
end = int(round(time.time() * 1000))
print(f'\tailia processing time {end - start} ms')
else:
output = detector.predict([img])
detect_object = post_processing(img, args.threshold, args.iou, output)
detect_object = reverse_letterbox(detect_object[0], org_img,
(IMAGE_HEIGHT, IMAGE_WIDTH))
# plot result
res_img = plot_results(detect_object, org_img, COCO_CATEGORY)
# plot result
cv2.imwrite(args.savepath, res_img)
print('Script finished successfully.')
def recognize_from_video(detector):
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
if args.detector:
img = cv2.cvtColor(frame, cv2.COLOR_BGR2BGRA)
detector.compute(img, args.threshold, args.iou)
res_img = plot_results(detector, frame, COCO_CATEGORY)
else:
img = letterbox_convert(frame, (IMAGE_HEIGHT, IMAGE_WIDTH))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img = np.transpose(img, [2, 0, 1])
img = img.astype(np.float32) / 255
img = np.expand_dims(img, 0)
output = detector.predict([img])
detect_object = post_processing(
img, args.threshold, args.iou, output
)
detect_object = reverse_letterbox(detect_object[0], frame, (IMAGE_HEIGHT,IMAGE_WIDTH))
res_img = plot_results(detect_object, frame, COCO_CATEGORY)
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, detect_object, 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_video():
# net initialize
detector = None
detector = ailia.Net(MODEL_PATH, WEIGHT_PATH, 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:
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 = letterbox_convert(frame, (IMAGE_HEIGHT, IMAGE_WIDTH))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img = np.transpose(img, [2, 0, 1])
img = img.astype(np.float32) / 255
img = np.expand_dims(img, 0)
output = detector.predict([img])
detect_object = yolov5_utils.post_processing(img, args.threshold,
args.iou, output)
detect_object = reverse_letterbox(detect_object[0], frame,
(IMAGE_HEIGHT, IMAGE_WIDTH))
res_img = plot_results(detect_object, frame, COCO_CATEGORY)
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()
print('Script finished successfully.')
def recognize_from_image():
# net initialize
detector = ailia.Net(MODEL_PATH, WEIGHT_PATH, env_id=args.env_id)
# input image loop
for image_path in args.input:
# prepare input data
logger.info(image_path)
# prepare input data
org_img = load_image(image_path)
org_img = cv2.cvtColor(org_img, cv2.COLOR_BGRA2BGR)
logger.info(f'input image shape: {org_img.shape}')
img = letterbox_convert(org_img, (IMAGE_HEIGHT, IMAGE_WIDTH))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img = np.transpose(img, [2, 0, 1])
img = img.astype(np.float32) / 255
img = np.expand_dims(img, 0)
# inference
logger.info('Start inference...')
if args.benchmark:
logger.info('BENCHMARK mode')
for i in range(5):
start = int(round(time.time() * 1000))
output = detector.predict([img])
end = int(round(time.time() * 1000))
logger.info(f'\tailia processing time {end - start} ms')
else:
output = detector.predict([img])
detect_object = yolov5_utils.post_processing(img, args.threshold,
args.iou, output)
detect_object = reverse_letterbox(detect_object[0], org_img,
(IMAGE_HEIGHT, IMAGE_WIDTH))
# plot result
res_img = plot_results(detect_object, org_img, COCO_CATEGORY)
# plot result
savepath = get_savepath(args.savepath, image_path)
logger.info(f'saved at : {savepath}')
cv2.imwrite(savepath, res_img)
logger.info('Script finished successfully.')
def compute_blazeface_with_keypoint(detector,
frame,
anchor_path='anchors.npy',
back=False,
min_score_thresh=DEFAULT_MIN_SCORE_THRESH):
if back:
BLAZEFACE_INPUT_IMAGE_HEIGHT = 256
BLAZEFACE_INPUT_IMAGE_WIDTH = 256
else:
BLAZEFACE_INPUT_IMAGE_HEIGHT = 128
BLAZEFACE_INPUT_IMAGE_WIDTH = 128
# preprocessing
image = letterbox_convert(
frame, (BLAZEFACE_INPUT_IMAGE_HEIGHT, BLAZEFACE_INPUT_IMAGE_WIDTH))
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image = image.transpose((2, 0, 1)) # channel first
image = image[np.newaxis, :, :, :] # (batch_size, channel, h, w)
input_data = image / 127.5 - 1.0
# inference
preds_ailia = detector.predict([input_data])
# postprocessing
face_detections = postprocess(preds_ailia,
anchor_path,
back=back,
min_score_thresh=min_score_thresh)
face_detections = face_detections[0]
detections = []
detections_eyes = []
for i, d in enumerate(face_detections):
# face position
obj = ailia.DetectorObject(category=0,
prob=1.0,
x=d[1],
y=d[0],
w=d[3] - d[1],
h=d[2] - d[0])
detections.append(obj)
# keypoints
obj = ailia.DetectorObject(category=0,
prob=1.0,
x=d[4],
y=d[5],
w=0,
h=0)
detections_eyes.append(obj)
obj = ailia.DetectorObject(category=0,
prob=1.0,
x=d[6],
y=d[7],
w=0,
h=0)
detections_eyes.append(obj)
# revert square from detections
detections = reverse_letterbox(
detections, frame,
(BLAZEFACE_INPUT_IMAGE_HEIGHT, BLAZEFACE_INPUT_IMAGE_WIDTH))
detections_eyes = reverse_letterbox(
detections_eyes, frame,
(BLAZEFACE_INPUT_IMAGE_HEIGHT, BLAZEFACE_INPUT_IMAGE_WIDTH))
# convert to keypoints
keypoints = []
for i in range(len(detections_eyes) // 2):
keypoint = {
"eye_left_x": detections_eyes[i * 2 + 0].x,
"eye_left_y": detections_eyes[i * 2 + 0].y,
"eye_right_x": detections_eyes[i * 2 + 1].x,
"eye_right_y": detections_eyes[i * 2 + 1].y
}
keypoints.append(keypoint)
return detections, keypoints
def recognize_from_image(detector):
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)
if not args.detector:
org_img = cv2.cvtColor(org_img, cv2.COLOR_BGRA2BGR)
logger.debug(f'input image shape: {org_img.shape}')
img = letterbox_convert(org_img, (IMAGE_HEIGHT, IMAGE_WIDTH))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img = np.transpose(img, [2, 0, 1])
img = img.astype(np.float32) / 255
img = np.expand_dims(img, 0)
# 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))
if args.detector:
detector.compute(org_img, args.threshold, args.iou)
else:
output = detector.predict([img])
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:
if args.detector:
detector.compute(org_img, args.threshold, args.iou)
else:
output = detector.predict([img])
if not args.detector:
detect_object = post_processing(img, args.threshold, args.iou, output)
detect_object = reverse_letterbox(detect_object[0], org_img, (IMAGE_HEIGHT,IMAGE_WIDTH))
res_img = plot_results(detect_object, org_img, COCO_CATEGORY)
else:
res_img = plot_results(detector, org_img, COCO_CATEGORY)
# plot result
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, detect_object, org_img, COCO_CATEGORY)
if args.profile:
print(detector.get_summary())
logger.info('Script finished successfully.')
