def create_tf_entry(label_and_data_info):
height = 800 # Image height
width = 1360 # Image width
filename = label_and_data_info[
0] # Filename of the image. Empty if image is not from file
img = Image.open(IMAGE_FOLDER + filename.decode())
b = io.BytesIO()
img.save(b, 'PNG')
encoded_image_data = b.getvalue() # Encoded image bytes
image_format = b'png' # b'jpeg' or b'png'
xmins = [
x / width for x in label_and_data_info[1]
] # List of normalized left x coordinates in bounding box (1 per box)
xmaxs = [x / width for x in label_and_data_info[2]
] # List of normalized right x coordinates in bounding box
# (1 per box)
ymins = [
y / height for y in label_and_data_info[3]
] # List of normalized top y coordinates in bounding box (1 per box)
ymaxs = [y / height for y in label_and_data_info[4]
] # List of normalized bottom y coordinates in bounding box
# (1 per box)
classes_text = label_and_data_info[
5] # List of string class name of bounding box (1 per box)
classes = label_and_data_info[
6] # List of integer class id of bounding box (1 per box)
tf_label_and_data = tf.train.Example(features=tf.train.Features(
feature={
'image/height':
dataset_util.int64_feature(height),
'image/width':
dataset_util.int64_feature(width),
'image/filename':
dataset_util.bytes_feature(filename),
'image/source_id':
dataset_util.bytes_feature(filename),
'image/encoded':
dataset_util.bytes_feature(encoded_image_data),
'image/format':
dataset_util.bytes_feature(image_format),
'image/object/bbox/xmin':
dataset_util.float_list_feature(xmins),
'image/object/bbox/xmax':
dataset_util.float_list_feature(xmaxs),
'image/object/bbox/ymin':
dataset_util.float_list_feature(ymins),
'image/object/bbox/ymax':
dataset_util.float_list_feature(ymaxs),
'image/object/class/text':
dataset_util.bytes_list_feature(classes_text),
'image/object/class/label':
dataset_util.int64_list_feature(classes),
}))
return tf_label_and_data
def create_tf_example(group, path):
with tf.gfile.GFile(os.path.join(path, '{}'.format(group.filename)),
'rb') as fid:
encoded_jpg = fid.read()
encoded_jpg_io = io.BytesIO(encoded_jpg)
image = Image.open(encoded_jpg_io)
width, height = image.size
filename = group.filename.encode('utf8')
image_format = b'jpg'
xmins = []
xmaxs = []
ymins = []
ymaxs = []
classes_text = []
classes = []
for index, row in group.object.iterrows():
xmins.append(row['xmin'] / width)
xmaxs.append(row['xmax'] / width)
ymins.append(row['ymin'] / height)
ymaxs.append(row['ymax'] / height)
classes_text.append(row['class'].encode('utf8'))
classes.append(class_text_to_int(row['class']))
tf_example = tf.train.Example(features=tf.train.Features(
feature={
'image/height':
dataset_util.int64_feature(height),
'image/width':
dataset_util.int64_feature(width),
'image/filename':
dataset_util.bytes_feature(filename),
'image/source_id':
dataset_util.bytes_feature(filename),
'image/encoded':
dataset_util.bytes_feature(encoded_jpg),
'image/format':
dataset_util.bytes_feature(image_format),
'image/object/bbox/xmin':
dataset_util.float_list_feature(xmins),
'image/object/bbox/xmax':
dataset_util.float_list_feature(xmaxs),
'image/object/bbox/ymin':
dataset_util.float_list_feature(ymins),
'image/object/bbox/ymax':
dataset_util.float_list_feature(ymaxs),
'image/object/class/text':
dataset_util.bytes_list_feature(classes_text),
'image/object/class/label':
dataset_util.int64_list_feature(classes),
}))
return tf_example
def make_tf_record(dataset_list, task_name, task_object, tfrecord_path):
tot = len(dataset_list)
old_percent, percent, iterator= 0, 0, 0
writer = tf.python_io.TFRecordWriter(tfrecord_path)
for dataset in dataset_list:
# 이미지 하나당 tf record feature 하나씩
img = storage.download_img_to_bytes(dataset['filename'])
img_format = dataset['filename'].split('.')[-1]
xmins, ymins, xmaxs, ymaxs, class_names, class_indices = [], [], [], [], [], []
for label in dataset[task_name]['label']:
xmins.append(label['x_min'])
ymins.append(label['y_min'])
xmaxs.append(label['x_max'])
ymaxs.append(label['y_max'])
# 사진 한장에 물체가 한가지 종류밖에 없음 만약
# tf record 파일 여러개인데 하나는 자동차, 다른 하나는 번호판이 되면
# class_indices 가 0 ,1 이 되야하는데 이걸 어떻게 구분하지? 여러개 다운로드로 해야할거 같기도 하고
# class_name = ['car', 'plate'] 라고 하면
# class_indices = ['0', '1'] 이런식으로 저장되야 car 랑 plate 두가지 물체가 서로 다르다라고 앎
class_indices.append(0)
class_names.append(task_object.encode('utf-8'))
feature = tf.train.Features(feature={
'image/height': dataset_util.int64_feature(dataset['height']),
'image/width': dataset_util.int64_feature(dataset['width']),
'image/filename': dataset_util.bytes_feature(dataset['filename'].encode('utf-8')),
'image/source_id': dataset_util.bytes_feature(dataset['filename'].encode('utf-8')),
'image/encoded': dataset_util.bytes_feature(img),
'image/format': dataset_util.bytes_feature(img_format.encode('utf-8')),
'image/object/bbox/xmin': dataset_util.float_list_feature(xmins),
'image/object/bbox/xmax': dataset_util.float_list_feature(xmaxs),
'image/object/bbox/ymin': dataset_util.float_list_feature(ymins),
'image/object/bbox/ymax': dataset_util.float_list_feature(ymaxs),
'image/object/class/text': dataset_util.bytes_list_feature(class_names),
'image/object/class/label': dataset_util.int64_list_feature(class_indices)
})
sample = tf.train.Example(features=feature)
writer.write(sample.SerializeToString())
# 확인용 코드
iterator += 1
percent = iterator * 100 // tot
if old_percent != percent:
print(percent)
old_percent = percent
writer.close()
def xml_to_tf(path_input, path_output):
xml_list = []
column_name = [
'filename', 'width', 'height', 'class', 'xmin', 'ymin', 'xmax', 'ymax'
]
writer = tf.io.TFRecordWriter(path_output)
files = os.listdir(path_input)
for file in files:
if file.endswith(".xml"):
xmlFile = path_input + file
tree = ET.parse(xmlFile)
root = tree.getroot()
filename = root[1].text
width = int(root[4][0].text)
height = int(root[4][1].text)
xmins = []
xmaxs = []
ymins = []
ymaxs = []
classes_text = []
classes = []
for member in root.findall('object'):
beer = member[0].text
xmin = int(member[4][0].text)
ymin = int(member[4][1].text)
xmax = int(member[4][2].text)
ymax = int(member[4][3].text)
xmins.append(xmin / width)
xmaxs.append(xmax / width)
ymins.append(ymin / height)
ymaxs.append(ymax / height)
classes_text.append(beer.encode('utf8'))
classes.append(class_text_to_int(beer))
with tf.io.gfile.GFile(
os.path.join(path_input, '{}'.format(filename)),
'rb') as fid:
encoded_jpg = fid.read()
tf_example = tf.train.Example(features=tf.train.Features(
feature={
'image/height':
dataset_util.int64_feature(height),
'image/width':
dataset_util.int64_feature(width),
'image/filename':
dataset_util.bytes_feature(filename.encode('utf8')),
'image/source_id':
dataset_util.bytes_feature(filename.encode('utf8')),
'image/encoded':
dataset_util.bytes_feature(encoded_jpg),
'image/format':
dataset_util.bytes_feature(IMAGE_FORMAT),
'image/object/bbox/xmin':
dataset_util.float_list_feature(xmins),
'image/object/bbox/xmax':
dataset_util.float_list_feature(xmaxs),
'image/object/bbox/ymin':
dataset_util.float_list_feature(ymins),
'image/object/bbox/ymax':
dataset_util.float_list_feature(ymaxs),
'image/object/class/text':
dataset_util.bytes_list_feature(classes_text),
'image/object/class/label':
dataset_util.int64_list_feature(classes),
}))
writer.write(tf_example.SerializeToString())
writer.close()
output_path = os.path.join(os.getcwd(), path_output)
print('Successfully created the TFRecords: {}'.format(output_path))
def create_tf_example(path, frame_num):
"""Creates a tf.Example proto for an individual image and its masks.
Args:
path: path to directory containing all images to be encoded
frame_num: The frame number of the image to be encoded.
Returns:
example: The created tf.Example.
"""
# TODO(user): Populate the following variables from your example.
img_path = path + '/' + str(frame_num) + '/image/image_' + str(
frame_num) + '.png'
img = imread(img_path)
# plt.imshow(img)
with tf.gfile.GFile(img_path, 'rb') as fid:
encoded_png = fid.read()
encoded_png_io = io.BytesIO(encoded_png)
# image = PIL.Image.open(encoded_png_io)
height = int(img.shape[0]) # Image height
width = int(img.shape[1]) # Image width
# print("Height: {}, width: {}".format(height, width))
filename = str(
frame_num
) # + '.png' # Filename of the image. Empty if image is not from file
# encoded_image_data = img.tobytes() # Encoded image bytes
image_format = b'png' # b'jpeg' or b'png'
xmins = [
] # List of normalized left x coordinates in bounding box (1 per box)
xmaxs = [] # List of normalized right x coordinates in bounding box
# (1 per box)
ymins = [
] # List of normalized top y coordinates in bounding box (1 per box)
ymaxs = [] # List of normalized bottom y coordinates in bounding box
# (1 per box)
classes_text = [] # List of string class name of bounding box (1 per box)
classes = [] # List of integer class id of bounding box (1 per box)
# masks = []
encoded_mask_png_list = []
for mask_id in (os.listdir(path + '/' + frame_num + '/masks/')):
# mask_img = imread(path + '/' + frame_num + '/masks/' + mask_id)
# mask_img = cv2.convertScaleAbs(mask_img)
# mask_img, contours, hierarchy = cv2.findContours(mask_img, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
# x, y, w, h = cv2.boundingRect(contours[0])
# mask_img = imread(path + '/' + frame_num + '/masks/' + mask_id)[:, :, :3]
# indices = np.where(mask_img)
mask_path = path + '/' + frame_num + '/masks/' + mask_id
with tf.gfile.GFile(mask_path, 'rb') as fid:
encoded_mask_png = fid.read()
encoded_png_io = io.BytesIO(encoded_mask_png)
mask = PIL.Image.open(encoded_png_io)
mask_np = np.asarray(mask)[:, :, :3]
# print(mask_np.shape)
indices = np.where(mask_np)
# print("Indices: ", indices)
if indices[0].size > 0:
xmin = float(np.min(indices[1]))
xmax = float(np.max(indices[1]))
ymin = float(np.min(indices[0]))
ymax = float(np.max(indices[0]))
xmins.append(xmin / width)
xmaxs.append(xmax / width)
ymins.append(ymin / height)
ymaxs.append(ymax / height)
classes_text.append('c_elegans'.encode('utf8'))
classes.append(int(1))
mask_remapped = mask_np.astype(np.uint8)
# masks.append(mask_remapped)
mask_img = PIL.Image.fromarray(mask_remapped)
# plt.imshow(mask_img)
# plt.show()
output = io.BytesIO()
mask_img.save(output, format='PNG')
encoded_mask_png_list.append(output.getvalue())
else:
pass
# encoded_mask_png_list = []
# for mask in masks:
# mask_img = PIL.Image.fromarray(mask)
# output = io.BytesIO()
# mask_img.save(output, format='PNG')
# encoded_mask_png_list.append(output.getvalue())
tf_example = tf.train.Example(features=tf.train.Features(
feature={
'image/height':
dataset_util.int64_feature(height),
'image/width':
dataset_util.int64_feature(width),
'image/filename':
dataset_util.bytes_feature(filename.encode('utf8')),
'image/source_id':
dataset_util.bytes_feature(filename.encode('utf8')),
'image/encoded':
dataset_util.bytes_feature(encoded_png),
'image/format':
dataset_util.bytes_feature(image_format),
'image/object/bbox/xmin':
dataset_util.float_list_feature(xmins),
'image/object/bbox/xmax':
dataset_util.float_list_feature(xmaxs),
'image/object/bbox/ymin':
dataset_util.float_list_feature(ymins),
'image/object/bbox/ymax':
dataset_util.float_list_feature(ymaxs),
'image/object/class/text':
dataset_util.bytes_list_feature(classes_text),
'image/object/class/label':
dataset_util.int64_list_feature(classes),
'image/object/mask':
dataset_util.bytes_list_feature(encoded_mask_png_list),
}))
return tf_example
def create_tf_example(image_path, image_name, height, width, sign_type_num,
x_min, y_min, x_max, y_max):
"""Creates a tf.Example proto from image.
Args:
example: The jpg encoded data of the cat image.
Returns:
example: The created tf.Example.
"""
# TODO(user): Populate the following variables from your example.
filename = image_name # Filename of the image. Empty if image is not from file
encoded_image_data = image_data = tf.gfile.FastGFile(image_path,
'rb').read()
image_format = b'jpeg' # b'jpeg' or b'png'
xmins = [
x_min / ORIGINAL_WIDTH
] # List of normalized left x coordinates in bounding box (1 per box)
xmaxs = [x_max / ORIGINAL_WIDTH
] # List of normalized right x coordinates in bounding box
# (1 per box)
ymins = [
y_min / ORIGINAL_HEIGHT
] # List of normalized top y coordinates in bounding box (1 per box)
ymaxs = [y_max / ORIGINAL_HEIGHT
] # List of normalized bottom y coordinates in bounding box
# (1 per box)
classes = [sign_type_num
] # List of integer class id of bounding box (1 per box)
classes_text = list(mydict.keys())[list(
mydict.values()).index(sign_type_num)]
print(
classes_text) # List of string class name of bounding box (1 per box)
tf_example = tf.train.Example(features=tf.train.Features(
feature={
'image/height':
dataset_util.int64_feature(height),
'image/width':
dataset_util.int64_feature(width),
'image/filename':
dataset_util.bytes_feature(filename.encode()),
'image/source_id':
dataset_util.bytes_feature(filename.encode()),
'image/encoded':
dataset_util.bytes_feature(encoded_image_data),
'image/format':
dataset_util.bytes_feature(image_format),
'image/object/bbox/xmin':
dataset_util.float_list_feature(xmins),
'image/object/bbox/xmax':
dataset_util.float_list_feature(xmaxs),
'image/object/bbox/ymin':
dataset_util.float_list_feature(ymins),
'image/object/bbox/ymax':
dataset_util.float_list_feature(ymaxs),
'image/object/class/text':
dataset_util.bytes_list_feature([classes_text.encode()]),
'image/object/class/label':
dataset_util.int64_list_feature(classes),
}))
return tf_example