def _dataset_parser(value):
"""Parse data to a fixed dimension input image and learning targets."""
with tf.name_scope('parser'):
data = example_decoder.decode(value)
source_id = data['source_id']
image = data['image']
boxes = data['groundtruth_boxes']
classes = data['groundtruth_classes']
classes = tf.reshape(tf.cast(classes, dtype=tf.float32), [-1, 1])
# the image normalization is identical to Cloud TPU ResNet-50
image = tf.image.convert_image_dtype(image, dtype=tf.float32)
image = _normalize_image(image)
if params['input_rand_hflip']:
image, boxes = preprocessor.random_horizontal_flip(image, boxes=boxes)
image_original_shape = tf.shape(image)
image, _ = preprocessor.resize_to_range(
image,
min_dimension=params['image_size'],
max_dimension=params['image_size'])
image_scale = tf.to_float(image_original_shape[0]) / tf.to_float(
tf.shape(image)[0])
image, boxes = preprocessor.scale_boxes_to_pixel_coordinates(
image, boxes, keypoints=None)
image = tf.image.pad_to_bounding_box(image, 0, 0, params['image_size'],
params['image_size'])
(cls_targets, box_targets,
num_positives) = anchor_labeler.label_anchors(boxes, classes)
source_id = tf.string_to_number(source_id, out_type=tf.float32)
row = (image, cls_targets, box_targets, num_positives, source_id,
image_scale)
return row
def _preprocess(example):
"""
example: a string tensor holding serialized tf example proto
"""
data = example_decoder.decode(example)
image = data["image"]
boxes = data["groundtruth_boxes"]
labels = data["groundtruth_classes"]
true_image_width = data["image/width"]
true_image_height = data["image/height"]
# augmentation
if is_training:
image = preprocessor.random_adjust_brightness(image)
image = preprocessor.random_adjust_saturation(image)
image = tf.image.convert_image_dtype(image, dtype=tf.float32)
image, boxes = preprocessor.random_horizontal_flip(image, boxes)
image, boxes, labels = preprocessor.random_crop_image(
image, boxes, labels)
else:
image = tf.image.convert_image_dtype(image, dtype=tf.float32)
# resize image
image = tf.image.resize_images(image, size=params["image_size"])
image = (image - 0.5) * 2
boxes = pad_to_fixed_size(boxes, [params["max_instance"], 4], -1)
labels = pad_to_fixed_size(labels, [params["max_instance"], 1], -1)
image_shape = tf.stack([true_image_height, true_image_width], axis=0)
return (image, boxes, labels, image_shape)
def random_horizontal_flip(self):
"""Randomly flip input image and bounding boxes."""
results = preprocessor.random_horizontal_flip(
self._image, boxes=self._boxes, masks=self._masks)
self._image = results[0]
self._boxes = results[1]
if self._masks is not None:
self._masks = results[2]
def _dataset_parser(value):
"""Parse data to a fixed dimension input image and learning targets."""
with tf.name_scope('parser'):
data = example_decoder.decode(value)
source_id = data['source_id']
# for xView dataset only; basically the original name is 122.tif and we will change it to number 122 later on.
# len = tf.size(tf.string_split([data['source_id']],""))
# source_id = tf.substr(data['source_id'],0,len - 4)
image = data['image']
boxes = data['groundtruth_boxes']
classes = data['groundtruth_classes']
classes = tf.reshape(tf.cast(classes, dtype=tf.float32), [-1, 1])
# Handle crowd annotations. As crowd annotations are not large
# instances, the model ignores them in training.
if params['skip_crowd']:
indices = tf.where(tf.logical_not(data['groundtruth_is_crowd']))
classes = tf.gather_nd(classes, indices)
boxes = tf.gather_nd(boxes, indices)
# the image normalization is identical to Cloud TPU ResNet-50
image = tf.image.convert_image_dtype(image, dtype=tf.float32)
image = _normalize_image(image)
if params['input_rand_hflip']:
image, boxes = preprocessor.random_horizontal_flip(image, boxes=boxes)
image_original_shape = tf.shape(image)
image, _ = preprocessor.resize_to_range(
image,
min_dimension=params['image_size'],
max_dimension=params['image_size'])
image_scale = tf.to_float(image_original_shape[0]) / tf.to_float(
tf.shape(image)[0])
image, boxes = preprocessor.scale_boxes_to_pixel_coordinates(
image, boxes, keypoints=None)
image = tf.image.pad_to_bounding_box(image, 0, 0, params['image_size'],
params['image_size'])
(cls_targets, box_targets,
num_positives) = anchor_labeler.label_anchors(boxes, classes)
#
# sess = tf.get_default_session()
# print("source id is", sess.run(source_id))
source_id = tf.string_to_number(source_id, out_type=tf.float32)
# sess = tf.get_default_session()
# print("after conversion, source id is", sess.run(source_id))
if params['use_bfloat16']:
image = tf.cast(image, dtype=tf.bfloat16)
row = (image, cls_targets, box_targets, num_positives, source_id,
image_scale)
return row
def random_horizontal_flip(image, boxes=None, masks=None):
"""Random horizontal flip the image, boxes, and masks.
Args:
image: a tensor of shape [height, width, 3] representing the image.
boxes: (Optional) a tensor of shape [num_boxes, 4] represneting the box
corners in normalized coordinates.
masks: (Optional) a tensor of shape [num_masks, height, width]
representing the object masks. Note that the size of the mask is the
same as the image.
Returns:
image: the processed image tensor after being randomly flipped.
boxes: None or the processed box tensor after being randomly flipped.
masks: None or the processed mask tensor after being randomly flipped.
"""
return preprocessor.random_horizontal_flip(image, boxes, masks)
def random_horizontal_flip(self):
"""Randomly flip input image and bounding boxes."""
self._image, self._boxes = preprocessor.random_horizontal_flip(
self._image, boxes=self._boxes)
def random_horizontal_flip(self):
"""Randomly flip input image and segmentation label."""
self._label = tf.expand_dims(self._label, 0)
self._image, self._label = preprocessor.random_horizontal_flip(
self._image, masks=self._label)
self._label = self._label[0, :, :]
def _parse_example(data):
with tf.name_scope('augmentation'):
source_id = data['source_id']
image = tf.image.convert_image_dtype(data['image'],
dtype=tf.float32)
raw_shape = tf.shape(image)
boxes = data['groundtruth_boxes']
classes = tf.reshape(data['groundtruth_classes'], [-1, 1])
# Only 80 of the 90 COCO classes are used.
class_map = tf.convert_to_tensor(ssd_constants.CLASS_MAP)
classes = tf.gather(class_map, classes)
classes = tf.cast(classes, dtype=tf.float32)
if self._is_training:
image, boxes, classes = ssd_crop(image, boxes, classes)
# random_horizontal_flip() is hard coded to flip with 50% chance.
mlperf_log.ssd_print(
key=mlperf_log.RANDOM_FLIP_PROBABILITY, value=0.5)
image, boxes = preprocessor.random_horizontal_flip(
image=image, boxes=boxes)
# TODO(shibow): Investigate the parameters for color jitter.
image = color_jitter(image,
brightness=0.125,
contrast=0.5,
saturation=0.5,
hue=0.05)
image = normalize_image(image)
if params['use_bfloat16']:
image = tf.cast(image, dtype=tf.bfloat16)
encoded_classes, encoded_boxes, num_matched_boxes = encode_labels(
boxes, classes)
# TODO(taylorrobie): Check that this cast is valid.
encoded_classes = tf.cast(encoded_classes, tf.int32)
labels = {
ssd_constants.NUM_MATCHED_BOXES: num_matched_boxes,
ssd_constants.BOXES: encoded_boxes,
ssd_constants.CLASSES: encoded_classes,
}
# This is for dataloader visualization; actual model doesn't use this.
if params['visualize_dataloader']:
box_coder = faster_rcnn_box_coder.FasterRcnnBoxCoder(
scale_factors=ssd_constants.BOX_CODER_SCALES)
decoded_boxes = tf.expand_dims(box_coder.decode(
rel_codes=tf.squeeze(encoded_boxes),
anchors=box_list.BoxList(
tf.convert_to_tensor(
DefaultBoxes()('ltrb')))).get(),
axis=0)
labels['decoded_boxes'] = tf.squeeze(decoded_boxes)
return image, labels
else:
mlperf_log.ssd_print(key=mlperf_log.INPUT_SIZE,
value=ssd_constants.IMAGE_SIZE)
image = tf.image.resize_images(
image[tf.newaxis, :, :, :],
size=(ssd_constants.IMAGE_SIZE,
ssd_constants.IMAGE_SIZE))[0, :, :, :]
image = normalize_image(image)
if params['use_bfloat16']:
image = tf.cast(image, dtype=tf.bfloat16)
def trim_and_pad(inp_tensor, dim_1):
"""Limit the number of boxes, and pad if necessary."""
inp_tensor = inp_tensor[:ssd_constants.
MAX_NUM_EVAL_BOXES]
num_pad = ssd_constants.MAX_NUM_EVAL_BOXES - tf.shape(
inp_tensor)[0]
inp_tensor = tf.pad(inp_tensor, [[0, num_pad], [0, 0]])
return tf.reshape(
inp_tensor,
[ssd_constants.MAX_NUM_EVAL_BOXES, dim_1])
boxes, classes = trim_and_pad(boxes,
4), trim_and_pad(classes, 1)
return {
ssd_constants.IMAGE:
image,
ssd_constants.BOXES:
boxes,
ssd_constants.CLASSES:
classes,
ssd_constants.SOURCE_ID:
tf.string_to_number(source_id, tf.int32),
ssd_constants.RAW_SHAPE:
raw_shape,
}
def _parse_example(data):
with tf.name_scope('augmentation'):
source_id = data['source_id']
image = data['image'] # dtype uint8
raw_shape = tf.shape(image)
boxes = data['groundtruth_boxes']
classes = tf.reshape(data['groundtruth_classes'], [-1, 1])
# Only 80 of the 90 COCO classes are used.
class_map = tf.convert_to_tensor(ssd_constants.CLASS_MAP)
classes = tf.gather(class_map, classes)
classes = tf.cast(classes, dtype=tf.float32)
if self._is_training:
image, boxes, classes = ssd_crop(image, boxes, classes)
# ssd_crop resizes and returns image of dtype float32 and does not
# change its range (i.e., value in between 0--255). Divide by 255.
# converts it to [0, 1] range. Not doing this before cropping to
# avoid dtype cast (which incurs additional memory copy).
image /= 255.0
# random_horizontal_flip() is hard coded to flip with 50% chance.
image, boxes = preprocessor.random_horizontal_flip(
image=image, boxes=boxes)
# TODO(shibow): Investigate the parameters for color jitter.
image = color_jitter(image,
brightness=0.125,
contrast=0.5,
saturation=0.5,
hue=0.05)
if params['use_bfloat16']:
image = tf.cast(image, dtype=tf.bfloat16)
encoded_classes, encoded_boxes, num_matched_boxes = encode_labels(
boxes, classes)
# TODO(taylorrobie): Check that this cast is valid.
encoded_classes = tf.cast(encoded_classes, tf.int32)
labels = {
ssd_constants.NUM_MATCHED_BOXES: num_matched_boxes,
ssd_constants.BOXES: encoded_boxes,
ssd_constants.CLASSES: tf.squeeze(encoded_classes,
axis=1),
}
# This is for dataloader visualization; actual model doesn't use this.
if params['visualize_dataloader']:
box_coder = faster_rcnn_box_coder.FasterRcnnBoxCoder(
scale_factors=ssd_constants.BOX_CODER_SCALES)
decoded_boxes = tf.expand_dims(box_coder.decode(
rel_codes=tf.squeeze(encoded_boxes),
anchors=box_list.BoxList(
tf.convert_to_tensor(
DefaultBoxes()('ltrb')))).get(),
axis=0)
labels['decoded_boxes'] = tf.squeeze(decoded_boxes)
return image, labels
else:
image = tf.image.resize_images(
image,
size=(ssd_constants.IMAGE_SIZE,
ssd_constants.IMAGE_SIZE))
# resize_image returns image of dtype float32 and does not change its
# range. Divide by 255 to convert image to [0, 1] range.
image /= 255.
if params['use_bfloat16']:
image = tf.cast(image, dtype=tf.bfloat16)
def trim_and_pad(inp_tensor, dim_1):
"""Limit the number of boxes, and pad if necessary."""
inp_tensor = inp_tensor[:ssd_constants.
MAX_NUM_EVAL_BOXES]
num_pad = ssd_constants.MAX_NUM_EVAL_BOXES - tf.shape(
inp_tensor)[0]
inp_tensor = tf.pad(inp_tensor, [[0, num_pad], [0, 0]])
return tf.reshape(
inp_tensor,
[ssd_constants.MAX_NUM_EVAL_BOXES, dim_1])
boxes, classes = trim_and_pad(boxes,
4), trim_and_pad(classes, 1)
sample = {
ssd_constants.IMAGE:
image,
ssd_constants.BOXES:
boxes,
ssd_constants.CLASSES:
classes,
ssd_constants.SOURCE_ID:
tf.string_to_number(source_id, tf.int32),
ssd_constants.RAW_SHAPE:
raw_shape,
}
if not self._is_training and self._count > params[
'eval_samples']:
sample[ssd_constants.IS_PADDED] = data[
ssd_constants.IS_PADDED]
return sample
def random_horizontal_flip(self):
"""Randomly flip input image and bounding boxes."""
self._label = tf.expand_dims(self._label, 0)
self._image, self._boxes, self._label = preprocessor.random_horizontal_flip(
self._image, boxes=self._boxes, masks=self._label)
self._label = self._label[0, :, :]