def resize_and_crop_boxes(self):
"""Resize boxes and crop it to the self._output dimension."""
boxlist = preprocessor.box_list.BoxList(self._boxes)
boxes = preprocessor.box_list_scale(boxlist, self._scaled_height,
self._scaled_width).get()
# Clip the boxes.
boxes = self.clip_boxes(boxes)
# Filter out ground truth boxes that are all zeros and corresponding classes
# and masks.
indices = tf.where(tf.not_equal(tf.reduce_sum(boxes, axis=1), 0))
boxes = tf.gather_nd(boxes, indices)
classes = tf.gather_nd(self._classes, indices)
self._masks = tf.gather_nd(self._masks, indices)
return boxes, classes
def _box_loss(box_outputs, box_targets, num_positives, delta=0.1):
"""Computes box regression loss."""
# delta is typically around the mean value of regression target.
# for instances, the regression targets of 512x512 input with 6 anchors on
# P3-P7 pyramid is about [0.1, 0.1, 0.2, 0.2].
normalizer = num_positives * 4.0
mask = tf.not_equal(box_targets, 0.0)
box_loss = tf.losses.huber_loss(
box_targets,
box_outputs,
weights=mask,
delta=delta,
reduction=tf.losses.Reduction.SUM)
box_loss /= normalizer
return box_loss
def resize_and_crop_boxes(self):
"""Resize boxes and crop it to the self._output dimension."""
boxlist = preprocessor.box_list.BoxList(self._boxes)
boxes = preprocessor.box_list_scale(
boxlist, self._scaled_height, self._scaled_width).get()
# Adjust box coordinates based on the offset.
box_offset = tf.stack([self._crop_offset_y, self._crop_offset_x,
self._crop_offset_y, self._crop_offset_x,])
boxes -= tf.to_float(tf.reshape(box_offset, [1, 4]))
# Clip the boxes.
boxes = self.clip_boxes(boxes)
# Filter out ground truth boxes that are all zeros.
indices = tf.where(tf.not_equal(tf.reduce_sum(boxes, axis=1), 0))
boxes = tf.gather_nd(boxes, indices)
classes = tf.gather_nd(self._classes, indices)
return boxes, classes
def label_anchors(self, gt_boxes, gt_labels):
"""Labels anchors with ground truth inputs.
Args:
gt_boxes: A float tensor with shape [N, 4] representing groundtruth boxes.
For each row, it stores [y0, x0, y1, x1] for four corners of a box.
gt_labels: A integer tensor with shape [N, 1] representing groundtruth
classes.
Returns:
cls_targets_dict: ordered dictionary with keys
[min_level, min_level+1, ..., max_level]. The values are tensor with
shape [height_l, width_l, num_anchors * num_classes]. The
height_l and width_l represent the dimension of class logits at l-th
level.
box_targets_dict: ordered dictionary with keys
[min_level, min_level+1, ..., max_level]. The values are tensor with
shape [height_l, width_l, num_anchors * 4]. The height_l and
width_l represent the dimension of bounding box regression output at
l-th level.
num_positives: scalar tensor storing number of positives in an image.
"""
gt_box_list = box_list.BoxList(gt_boxes)
anchor_box_list = box_list.BoxList(self._anchors.boxes)
# cls_weights, box_weights are not used
cls_targets, _, box_targets, _, matches = self._target_assigner.assign(
anchor_box_list, gt_box_list, gt_labels)
# class labels start from 1 and the background class = -1
cls_targets -= 1
# create one-hot labels
cls_targets_one_hot = tf.one_hot(tf.cast(cls_targets, dtype=tf.int32),
self._num_classes)
cls_targets_one_hot = tf.reshape(cls_targets_one_hot,
[-1, self._num_classes])
cls_targets_dict = self._unpack_labels(cls_targets_one_hot)
box_targets_dict = self._unpack_labels(box_targets)
num_positives = tf.reduce_sum(
tf.cast(tf.not_equal(matches.match_results, -1), tf.float32))
return cls_targets_dict, box_targets_dict, num_positives