with tf.Session(config=config) as sess:
init_op = tf.global_variables_initializer()
model = sess.run(init_op)
if os.path.isfile(os.getcwd() + "/" + cfg.weights_dir + "/checkpoint"):
saver.restore(sess, model_file)
print("Restored model")
yolo.set_training(False)
anchors = np.reshape(np.array(cfg.anchors), [-1, 2])
images = load_file(sys.argv[1:])
#normalise data between 0 and 1
imgs = (np.array([row[0] for row in images])/255)
boxes = sess.run(yolo.output, feed_dict={
yolo.x: imgs,
yolo.anchors: anchors,
})
proc_boxes = yolo.convert_net_to_bb(boxes, filter_top=True)
for box in proc_boxes:
cls = yolo.names[int(box[0])]
hex = cls.encode('utf-8').hex()[0:6]
color = tuple(int(hex[k:k+2], 16) for k in (0, 2 ,4))
print(cls, box[1], box[2], box[3], box[4], box[5])
if len(v_labels) == 0:
continue
res, correct, iou = sess.run(
[yolo.output, yolo.matches, yolo.best_iou],
feed_dict={
yolo.train_bounding_boxes: v_labels,
yolo.train_object_recognition: v_obj_detection,
yolo.x: v_imgs,
yolo.anchors: anchors,
yolo.iou_threshold: iou_threshold,
yolo.object_detection_threshold:
confidence_threshold
})
labels = yolo.convert_net_to_bb(res, filter_top=True)[0]
#v_obj_detection = np.zeros_like(v_obj_detection)
o_img, o_h, o_w, = res.shape[:3]
img = o_img - 1
if one_class:
v_obj_detection = np.zeros_like(v_obj_detection)
while img >= 0:
h = o_h - 1
while h >= 0:
w = o_w - 1
while w >= 0:
