def inference(test_dir, inference_save_path):
test_imgname_list = [
os.path.join(test_dir, img_name) for img_name in os.listdir(test_dir)
if img_name.endswith(('.jpg', '.png', '.jpeg', '.tif', '.tiff'))
]
assert len(test_imgname_list) != 0, 'test_dir has no imgs there.' \
' Note that, we only support img format of (.jpg, .png, and .tiff) '
faster_rcnn = build_whole_network_r3det.DetectionNetwork(
base_network_name=cfgs.NET_NAME, is_training=False)
detect(det_net=faster_rcnn,
inference_save_path=inference_save_path,
real_test_imgname_list=test_imgname_list)
def eval(num_imgs, args):
txt_name = '{}.txt'.format(cfgs.VERSION)
if not args.show_box:
if not os.path.exists(txt_name):
fw = open(txt_name, 'w')
fw.close()
fr = open(txt_name, 'r')
img_filter = fr.readlines()
print('****************************' * 3)
print('Already tested imgs:', img_filter)
print('****************************' * 3)
fr.close()
test_imgname_list = [
os.path.join(args.test_dir, img_name)
for img_name in os.listdir(args.test_dir)
if img_name.endswith(('.jpg', '.png', '.jpeg', '.tif',
'.tiff')) and (img_name +
'\n' not in img_filter)
]
else:
test_imgname_list = [
os.path.join(args.test_dir, img_name)
for img_name in os.listdir(args.test_dir)
if img_name.endswith(('.jpg', '.png', '.jpeg', '.tif', '.tiff'))
]
assert len(test_imgname_list) != 0, 'test_dir has no imgs there.' \
' Note that, we only support img format of (.jpg, .png, and .tiff) '
if num_imgs == np.inf:
real_test_img_list = test_imgname_list
else:
real_test_img_list = test_imgname_list[:num_imgs]
retinanet = build_whole_network_r3det.DetectionNetwork(
base_network_name=cfgs.NET_NAME, is_training=False)
test_dota(det_net=retinanet,
real_test_img_list=real_test_img_list,
args=args,
txt_name=txt_name)
if not args.show_box:
os.remove(txt_name)
def eval(num_imgs, img_dir, image_ext, test_annotation_path, draw_imgs):
r3det = build_whole_network_r3det.DetectionNetwork(base_network_name=cfgs.NET_NAME,
is_training=False)
all_boxes_r = eval_with_plac(img_dir=img_dir, det_net=r3det,
num_imgs=num_imgs, image_ext=image_ext, draw_imgs=draw_imgs)
# with open(cfgs.VERSION + '_detections_r.pkl', 'rb') as f2:
# all_boxes_r = pickle.load(f2)
#
# print(len(all_boxes_r))
imgs = os.listdir(img_dir)
real_test_imgname_list = [i.split(image_ext)[0] for i in imgs]
print(10 * "**")
print('rotation eval:')
voc_eval_r.voc_evaluate_detections(all_boxes=all_boxes_r,
test_imgid_list=real_test_imgname_list,
test_annotation_path=test_annotation_path)
def train():
with tf.Graph().as_default(), tf.device('/cpu:0'):
num_gpu = len(cfgs.GPU_GROUP.strip().split(','))
global_step = slim.get_or_create_global_step()
lr = warmup_lr(cfgs.LR, global_step, cfgs.WARM_SETP, num_gpu)
tf.summary.scalar('lr', lr)
with tf.name_scope('get_batch'):
if cfgs.IMAGE_PYRAMID:
shortside_len_list = tf.constant(cfgs.IMG_SHORT_SIDE_LEN)
shortside_len = tf.random_shuffle(shortside_len_list)[0]
else:
shortside_len = cfgs.IMG_SHORT_SIDE_LEN
img_name_batch, img_batch, gtboxes_and_label_batch, num_objects_batch, img_h_batch, img_w_batch = \
next_batch(dataset_name=cfgs.DATASET_NAME,
batch_size=cfgs.BATCH_SIZE * num_gpu,
shortside_len=shortside_len,
is_training=True)
optimizer = tf.train.MomentumOptimizer(lr, momentum=cfgs.MOMENTUM)
r3det = build_whole_network_r3det.DetectionNetwork(
base_network_name=cfgs.NET_NAME, is_training=True)
# data processing
inputs_list = []
for i in range(num_gpu):
img = tf.expand_dims(img_batch[i], axis=0)
if cfgs.NET_NAME in [
'resnet152_v1d', 'resnet101_v1d', 'resnet50_v1d'
]:
img = img / tf.constant([cfgs.PIXEL_STD])
gtboxes_and_label_r = tf.py_func(backward_convert,
inp=[gtboxes_and_label_batch[i]],
Tout=tf.float32)
gtboxes_and_label_r = tf.reshape(gtboxes_and_label_r, [-1, 6])
gtboxes_and_label_h = get_horizen_minAreaRectangle(
gtboxes_and_label_batch[i])
gtboxes_and_label_h = tf.reshape(gtboxes_and_label_h, [-1, 5])
num_objects = num_objects_batch[i]
num_objects = tf.cast(tf.reshape(num_objects, [
-1,
]), tf.float32)
img_h = img_h_batch[i]
img_w = img_w_batch[i]
inputs_list.append([
img, gtboxes_and_label_h, gtboxes_and_label_r, num_objects,
img_h, img_w
])
tower_grads = []
biases_regularizer = tf.no_regularizer
weights_regularizer = tf.contrib.layers.l2_regularizer(
cfgs.WEIGHT_DECAY)
total_loss_dict = {
'cls_loss': tf.constant(0., tf.float32),
'reg_loss': tf.constant(0., tf.float32),
'refine_cls_loss': tf.constant(0., tf.float32),
'refine_reg_loss': tf.constant(0., tf.float32),
'total_losses': tf.constant(0., tf.float32),
}
if cfgs.USE_SUPERVISED_MASK:
total_loss_dict['mask_loss'] = tf.constant(0., tf.float32)
with tf.variable_scope(tf.get_variable_scope()):
for i in range(num_gpu):
with tf.device('/gpu:%d' % i):
with tf.name_scope('tower_%d' % i):
with slim.arg_scope(
[slim.model_variable, slim.variable],
device='/device:CPU:0'):
with slim.arg_scope(
[
slim.conv2d, slim.conv2d_in_plane,
slim.conv2d_transpose,
slim.separable_conv2d, slim.fully_connected
],
weights_regularizer=weights_regularizer,
biases_regularizer=biases_regularizer,
biases_initializer=tf.constant_initializer(
0.0)):
gtboxes_and_label_h, gtboxes_and_label_r = tf.py_func(
get_gtboxes_and_label,
inp=[
inputs_list[i][1], inputs_list[i][2],
inputs_list[i][3]
],
Tout=[tf.float32, tf.float32])
gtboxes_and_label_h = tf.reshape(
gtboxes_and_label_h, [-1, 5])
gtboxes_and_label_r = tf.reshape(
gtboxes_and_label_r, [-1, 6])
img = inputs_list[i][0]
img_shape = inputs_list[i][-2:]
img = tf.image.crop_to_bounding_box(
image=img,
offset_height=0,
offset_width=0,
target_height=tf.cast(
img_shape[0], tf.int32),
target_width=tf.cast(
img_shape[1], tf.int32))
outputs = r3det.build_whole_detection_network(
input_img_batch=img,
gtboxes_batch_h=gtboxes_and_label_h,
gtboxes_batch_r=gtboxes_and_label_r,
gpu_id=i)
gtboxes_in_img_h = draw_boxes_with_categories(
img_batch=img,
boxes=gtboxes_and_label_h[:, :-1],
labels=gtboxes_and_label_h[:, -1],
method=0)
gtboxes_in_img_r = draw_boxes_with_categories(
img_batch=img,
boxes=gtboxes_and_label_r[:, :-1],
labels=gtboxes_and_label_r[:, -1],
method=1)
tf.summary.image(
'Compare/gtboxes_h_gpu:%d' % i,
gtboxes_in_img_h)
tf.summary.image(
'Compare/gtboxes_r_gpu:%d' % i,
gtboxes_in_img_r)
if cfgs.ADD_BOX_IN_TENSORBOARD:
detections_in_img = draw_boxes_with_categories_and_scores(
img_batch=img,
boxes=outputs[0],
scores=outputs[1],
labels=outputs[2],
method=1)
tf.summary.image(
'Compare/final_detection_gpu:%d' % i,
detections_in_img)
loss_dict = outputs[-1]
total_losses = 0.0
for k in loss_dict.keys():
total_losses += loss_dict[k]
total_loss_dict[
k] += loss_dict[k] / num_gpu
total_losses /= num_gpu
total_loss_dict['total_losses'] += total_losses
if i == num_gpu - 1:
regularization_losses = tf.get_collection(
tf.GraphKeys.REGULARIZATION_LOSSES)
# weight_decay_loss = tf.add_n(slim.losses.get_regularization_losses())
total_losses = total_losses + tf.add_n(
regularization_losses)
tf.get_variable_scope().reuse_variables()
grads = optimizer.compute_gradients(total_losses)
if cfgs.GRADIENT_CLIPPING_BY_NORM is not None:
grads = slim.learning.clip_gradient_norms(
grads, cfgs.GRADIENT_CLIPPING_BY_NORM)
tower_grads.append(grads)
for k in total_loss_dict.keys():
tf.summary.scalar('{}/{}'.format(k.split('_')[0], k),
total_loss_dict[k])
if len(tower_grads) > 1:
grads = sum_gradients(tower_grads)
else:
grads = tower_grads[0]
if cfgs.MUTILPY_BIAS_GRADIENT is not None:
final_gvs = []
with tf.variable_scope('Gradient_Mult'):
for grad, var in grads:
scale = 1.
if '/biases:' in var.name:
scale *= cfgs.MUTILPY_BIAS_GRADIENT
if 'conv_new' in var.name:
scale *= 3.
if not np.allclose(scale, 1.0):
grad = tf.multiply(grad, scale)
final_gvs.append((grad, var))
apply_gradient_op = optimizer.apply_gradients(
final_gvs, global_step=global_step)
else:
apply_gradient_op = optimizer.apply_gradients(
grads, global_step=global_step)
variable_averages = tf.train.ExponentialMovingAverage(
0.9999, global_step)
variables_averages_op = variable_averages.apply(
tf.trainable_variables())
train_op = tf.group(apply_gradient_op, variables_averages_op)
# train_op = optimizer.apply_gradients(final_gvs, global_step=global_step)
summary_op = tf.summary.merge_all()
restorer, restore_ckpt = r3det.get_restorer()
saver = tf.train.Saver(max_to_keep=5)
init_op = tf.group(tf.global_variables_initializer(),
tf.local_variables_initializer())
tfconfig = tf.ConfigProto(allow_soft_placement=True,
log_device_placement=False)
tfconfig.gpu_options.allow_growth = True
with tf.Session(config=tfconfig) as sess:
sess.run(init_op)
# sess.run(tf.initialize_all_variables())
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(coord=coord, sess=sess)
summary_path = os.path.join(cfgs.SUMMARY_PATH, cfgs.VERSION)
tools.mkdir(summary_path)
summary_writer = tf.summary.FileWriter(summary_path,
graph=sess.graph)
if not restorer is None:
restorer.restore(sess, restore_ckpt)
print('restore model')
for step in range(cfgs.MAX_ITERATION // num_gpu):
training_time = time.strftime('%Y-%m-%d %H:%M:%S',
time.localtime(time.time()))
if step % cfgs.SHOW_TRAIN_INFO_INTE != 0 and step % cfgs.SMRY_ITER != 0:
_, global_stepnp = sess.run([train_op, global_step])
else:
if step % cfgs.SHOW_TRAIN_INFO_INTE == 0 and step % cfgs.SMRY_ITER != 0:
start = time.time()
_, global_stepnp, total_loss_dict_ = \
sess.run([train_op, global_step, total_loss_dict])
end = time.time()
print('***' * 20)
print("""%s: global_step:%d current_step:%d""" %
(training_time,
(global_stepnp - 1) * num_gpu, step * num_gpu))
print("""per_cost_time:%.3fs""" %
((end - start) / num_gpu))
loss_str = ''
for k in total_loss_dict_.keys():
loss_str += '%s:%.3f\n' % (k, total_loss_dict_[k])
print(loss_str)
if np.isnan(total_loss_dict_['total_losses']):
sys.exit(0)
else:
if step % cfgs.SMRY_ITER == 0:
_, global_stepnp, summary_str = sess.run(
[train_op, global_step, summary_op])
summary_writer.add_summary(
summary_str, (global_stepnp - 1) * num_gpu)
summary_writer.flush()
if (step > 0 and step % (cfgs.SAVE_WEIGHTS_INTE // num_gpu)
== 0) or (step >= cfgs.MAX_ITERATION // num_gpu - 1):
save_dir = os.path.join(cfgs.TRAINED_CKPT, cfgs.VERSION)
if not os.path.exists(save_dir):
os.mkdir(save_dir)
save_ckpt = os.path.join(
save_dir, '{}_'.format(cfgs.DATASET_NAME) + str(
(global_stepnp - 1) * num_gpu) + 'model.ckpt')
saver.save(sess, save_ckpt)
print(' weights had been saved')
coord.request_stop()
coord.join(threads)
def build_detection_graph():
# 1. preprocess img
img_plac = tf.placeholder(dtype=tf.float32, shape=[1, 640, 640, 3],
name='input_img') # is RGB. not BGR
#img_plac = tf.placeholder(dtype=tf.uint8, shape=[1, None, None, 3],
# name='input_img') # is RGB. not BGR
#raw_shape = tf.shape(img_plac)
#raw_h, raw_w = tf.to_float(raw_shape[0]), tf.to_float(raw_shape[1])
#img_batch = tf.cast(img_plac, tf.float32)
#img_plac = tf.placeholder(dtype=tf.uint8, shape=[None, None, 3],
# name='input_img') # is RGB. not GBR
#raw_shape = tf.shape(img_plac)
#raw_h, raw_w = tf.to_float(raw_shape[0]), tf.to_float(raw_shape[1])
#img_batch = tf.cast(img_plac, tf.float32)
#img_batch = short_side_resize_for_inference_data(img_tensor=img_batch,
# target_shortside_len=cfgs.IMG_SHORT_SIDE_LEN,
# length_limitation=cfgs.IMG_MAX_LENGTH)
#if cfgs.NET_NAME in ['resnet152_v1d', 'resnet101_v1d', 'resnet50_v1d']:
# img_batch = (img_batch / 255 - tf.constant(cfgs.PIXEL_MEAN_)) / tf.constant(cfgs.PIXEL_STD)
#else:
# img_batch = img_batch - tf.constant(cfgs.PIXEL_MEAN)
#img_batch = tf.expand_dims(img_batch, axis=0)
#img_batch = img_batch - tf.constant(cfgs.PIXEL_MEAN)
#img_batch = tf.expand_dims(img_batch, axis=0) # [1, None, None, 3]
det_net = build_whole_network_r3det.DetectionNetwork(base_network_name=cfgs.NET_NAME,
is_training=False)
#det_net = build_whole_network.DetectionNetwork(base_network_name=cfgs.NET_NAME,
# is_training=False)
detection_boxes, detection_scores, detection_category = det_net.build_whole_detection_network(
input_img_batch=img_plac,
gtboxes_batch_h=None,
gtboxes_batch_r=None)
#detection_boxes, detection_scores, detection_category = det_net.build_whole_detection_network(
# input_img_batch=img_batch,
# gtboxes_batch_h=None,
# gtboxes_batch_r=None)
#detected_boxes, detection_scores, detection_category = det_net.build_whole_detection_network(
# input_img_batch=img_batch,
# gtboxes_batch=None)
#x_c, y_c, w, h, theta = detection_boxes[:, 0], detection_boxes[:, 1],\
# detection_boxes[:, 2], detection_boxes[:, 3],\
# detection_boxes[:, 4]
#xmin, ymin, xmax, ymax = detected_boxes[:, 0], detected_boxes[:, 1], \
# detected_boxes[:, 2], detected_boxes[:, 3]
#resized_shape = tf.shape(img_batch)
#resized_h, resized_w = tf.to_float(resized_shape[1]), tf.to_float(resized_shape[2])
#x_c = x_c * raw_w / resized_w
#w = w * raw_w / resized_w
#xmin = xmin * raw_w / resized_w
#xmax = xmax * raw_w / resized_w
#y_c = y_c * raw_h / resized_h
#h = h * raw_h / resized_h
#ymin = ymin * raw_h / resized_h
#ymax = ymax * raw_h / resized_h
boxes = tf.transpose(tf.stack([detection_boxes[:, 0], detection_boxes[:, 1],
detection_boxes[:, 2], detection_boxes[:, 3],
detection_boxes[:, 4]]))
#boxes = tf.transpose(tf.stack([x_c, y_c, w, h, theta]))
#boxes = tf.transpose(tf.stack([xmin, ymin, xmax, ymax]))
dets = tf.concat([tf.reshape(detection_category, [-1, 1]),
tf.reshape(detection_scores, [-1, 1]),
boxes], axis=1, name='DetResults')
return dets
