x = x.reshape(-1)
e_x = np.exp(x - np.max(x))
return e_x / e_x.sum(axis=0)
def postprocess(scores):
"""This function takes the scores generated by the network and
returns the class IDs in decreasing order of probability."""
prob = softmax(scores)
prob = np.squeeze(prob)
classes = np.argsort(prob)[::-1]
return classes
# Launch Zetane
zcontext = ztn.Context().launch()
zonnx = zcontext.model()
zimg_input = zcontext.image()
# provide the name of the model and width and height of the images for the variables
model = directory + r'/emotion-ferplus-8.onnx'
width = 64
height = 64
filepath = directory + r'/happy.jpg'
preprocess_image = preprocess(filepath)
x = np.array(preprocess_image).astype('float32')
#np.save('testferemotions.npy', x)
zimg_input.position(2, 7, 0).scale(.5, .5, .5).rotation(0, -1.57, 0).update(
filepath=filepath) #.rotation(0, 0, -1.57).
img_path = os.path.join(dir_path, 'pic.jpg')
color = (255, 255, 255)
orig_image = cv2.imread(img_path)
boxes, labels, probs = faceDetector(orig_image)
box = scale(boxes[0, :])
cropped = cropImage(orig_image, box)
age = ageClassifier(cropped)
cv2.rectangle(orig_image, (box[0], box[1]), (box[2], box[3]), color, 8)
cv2.putText(orig_image, f'{age}', (box[0], box[1] - 10),
cv2.FONT_HERSHEY_SIMPLEX, 6, color, 8, cv2.LINE_AA)
output_path = os.path.join(dir_path, 'output.jpg')
cv2.imwrite(output_path, orig_image)
zcontext = ztn.Context()
zcontext.launch()
zcontext.clear_universe()
zmodel = zcontext.model()
zimg_1 = zcontext.image()
zimg_2 = zcontext.image()
image = cv2.cvtColor(cropped, cv2.COLOR_BGR2RGB)
image = cv2.resize(image, (224, 224))
image = np.transpose(image, [2, 0, 1])
image = np.expand_dims(image, axis=0)
image = image.astype(np.float32)
zmodel.onnx(age_classifier_onnx).update()
zmodel.inputs(inputs=image).update()
def __init__(self):
self.context = ztn.Context()
self.iimage = self.context.image().position(-10.0, 2.5, 0.0).scale(
0.25, 0.25, 0.0)
self.vmodel = self.context.model()
self.context.clear_universe()
def main(_argv):
physical_devices = tf.config.experimental.list_physical_devices('GPU')
for physical_device in physical_devices:
tf.config.experimental.set_memory_growth(physical_device, True)
if FLAGS.tiny:
yolo = YoloV3Tiny(classes=FLAGS.num_classes)
else:
yolo = YoloV3(classes=FLAGS.num_classes)
yolo.load_weights(FLAGS.weights)
logging.info('weights loaded')
class_names = [c.strip() for c in open(FLAGS.classes).readlines()]
logging.info('classes loaded')
times = []
try:
vid = cv2.VideoCapture(int(FLAGS.video))
except:
vid = cv2.VideoCapture(FLAGS.video)
out = None
ctxt = zetane.Context()
ctxt.clear_universe()
input_panel, output_panel = make_io_panels(ctxt)
counter = 0
if FLAGS.output:
# by default VideoCapture returns float instead of int
width = int(vid.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(vid.get(cv2.CAP_PROP_FRAME_HEIGHT))
fps = int(vid.get(cv2.CAP_PROP_FPS))
codec = cv2.VideoWriter_fourcc(*FLAGS.output_format)
out = cv2.VideoWriter(FLAGS.output, codec, fps, (width, height))
while True:
_, img = vid.read()
if img is None:
logging.warning("Empty Frame")
time.sleep(0.1)
continue
img_in = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img_in = tf.expand_dims(img_in, 0)
img_in = transform_images(img_in, FLAGS.size)
image_np = np.transpose(img_in.numpy(), (1, 2, 3, 0))
if counter == 0:
to_fit = 0.15 / image_np.shape[2]
zinput = ctxt.image().data(image_np).scale(to_fit, to_fit).send_to(input_panel).update()
zmodel = ctxt.model().keras(yolo).inputs(img_in.numpy()).update()
else:
zinput.data(image_np).update()
zmodel.inputs(img_in.numpy()).update()
t1 = time.time()
#boxes, scores, classes, nums = yolo.predict(img_in)
bbox, confidence, class_probs, scores = yolo(img_in)
boxes, scores, classes, nums = tf.image.combined_non_max_suppression(
boxes=tf.reshape(bbox, (tf.shape(bbox)[0], -1, 1, 4)),
scores=tf.reshape(
scores, (tf.shape(scores)[0], -1, tf.shape(scores)[-1])),
max_output_size_per_class=FLAGS.yolo_max_boxes,
max_total_size=FLAGS.yolo_max_boxes,
iou_threshold=FLAGS.yolo_iou_threshold,
score_threshold=FLAGS.yolo_score_threshold
)
t2 = time.time()
times.append(t2-t1)
times = times[-20:]
out_img = draw_outputs(img/255.0, (boxes, scores, classes, nums), class_names)
out_img = cv2.putText(out_img, "Time: {:.2f}ms".format(sum(times)/len(times)*1000), (0, 30), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (0, 0, 255), 2)
if counter == 0:
to_fit = 0.15 / out_img.shape[2]
zoutput = ctxt.image().data(out_img).scale(to_fit, to_fit).send_to(output_panel).update()
else:
zoutput.data(out_img).debug();
#if FLAGS.output:
# out.write(img)
#cv2.imshow('output', img)
counter += 1
if cv2.waitKey(1) == ord('q'):
break
ctxt.disconnect()
cv2.destroyAllWindows()
def main(_argv):
physical_devices = tf.config.experimental.list_physical_devices('GPU')
for physical_device in physical_devices:
tf.config.experimental.set_memory_growth(physical_device, True)
if FLAGS.tiny:
yolo = YoloV3Tiny(classes=FLAGS.num_classes)
else:
yolo = YoloV3(classes=FLAGS.num_classes)
yolo.load_weights(FLAGS.weights).expect_partial()
logging.info('weights loaded')
class_names = [c.strip() for c in open(FLAGS.classes).readlines()]
logging.info('classes loaded')
if FLAGS.tfrecord:
dataset = load_tfrecord_dataset(
FLAGS.tfrecord, FLAGS.classes, FLAGS.size)
dataset = dataset.shuffle(512)
img_raw, _label = next(iter(dataset.take(1)))
else:
img_raw = tf.image.decode_image(
open(FLAGS.image, 'rb').read(), channels=3)
img = tf.expand_dims(img_raw, 0)
img = transform_images(img, FLAGS.size)
ctxt = zetane.Context()
ctxt.clear_universe()
input_panel, output_panel = make_io_panels(ctxt)
image_np = np.transpose(img.numpy(), (1, 2, 3, 0))
to_fit = 0.15 / image_np.shape[2]
zinput = ctxt.image().data(image_np).scale(to_fit, to_fit).send_to(input_panel).update()
zmodel = ctxt.model().keras(yolo).inputs(img.numpy()).update()
t1 = time.time()
#boxes, scores, classes, nums = yolo(img)
bbox, confidence, class_probs, scores = yolo(img)
boxes, scores, classes, nums = tf.image.combined_non_max_suppression(
boxes=tf.reshape(bbox, (tf.shape(bbox)[0], -1, 1, 4)),
scores=tf.reshape(
scores, (tf.shape(scores)[0], -1, tf.shape(scores)[-1])),
max_output_size_per_class=FLAGS.yolo_max_boxes,
max_total_size=FLAGS.yolo_max_boxes,
iou_threshold=FLAGS.yolo_iou_threshold,
score_threshold=FLAGS.yolo_score_threshold
)
t2 = time.time()
logging.info('time: {}'.format(t2 - t1))
logging.info('detections:')
for i in range(nums[0]):
logging.info('\t{}, {}, {}'.format(class_names[int(classes[0][i])],
np.array(scores[0][i]),
np.array(boxes[0][i])))
#out_img = cv2.cvtColor(img_raw.numpy(), cv2.COLOR_RGB2BGR)
out_img = draw_outputs(img_raw.numpy()/225.0, (boxes, scores, classes, nums), class_names)
to_fit = 0.15 / out_img.shape[2]
zoutput = ctxt.image().data(out_img).scale(to_fit, to_fit).send_to(output_panel).update()
#cv2.imwrite(FLAGS.output, img)
#logging.info('output saved to: {}'.format(FLAGS.output))
ctxt.disconnect()
