def test_with_no_scores_field(self):
box_mask_list = np_box_mask_list.BoxMaskList(
box_data=self.boxes1, mask_data=self.masks1)
max_output_size = 3
iou_threshold = 0.5
with self.assertRaises(ValueError):
np_box_mask_list_ops.non_max_suppression(
box_mask_list, max_output_size, iou_threshold)
def test_with_no_scores_field(self):
box_mask_list = np_box_mask_list.BoxMaskList(
box_data=self.boxes1, mask_data=self.masks1)
max_output_size = 3
iou_threshold = 0.5
with self.assertRaises(ValueError):
np_box_mask_list_ops.non_max_suppression(
box_mask_list, max_output_size, iou_threshold)
def _get_overlaps_and_scores_mask_mode(self, detected_boxes,
detected_scores, detected_masks,
groundtruth_boxes,
groundtruth_masks,
groundtruth_is_group_of_list):
"""Computes overlaps and scores between detected and groudntruth masks.
Args:
detected_boxes: A numpy array of shape [N, 4] representing detected box
coordinates
detected_scores: A 1-d numpy array of length N representing classification
score
detected_masks: A uint8 numpy array of shape [N, height, width]. If not
None, the scores will be computed based on masks.
groundtruth_boxes: A numpy array of shape [M, 4] representing ground truth
box coordinates
groundtruth_masks: A uint8 numpy array of shape [M, height, width].
groundtruth_is_group_of_list: A boolean numpy array of length M denoting
whether a ground truth box has group-of tag. If a groundtruth box is
group-of box, every detection matching this box is ignored.
Returns:
iou: A float numpy array of size [num_detected_boxes, num_gt_boxes]. If
gt_non_group_of_boxlist.num_boxes() == 0 it will be None.
ioa: A float numpy array of size [num_detected_boxes, num_gt_boxes]. If
gt_group_of_boxlist.num_boxes() == 0 it will be None.
scores: The score of the detected boxlist.
num_boxes: Number of non-maximum suppressed detected boxes.
"""
#print('len(detected_boxes)',len(detected_boxes))
detected_boxlist = np_box_mask_list.BoxMaskList(
box_data=detected_boxes, mask_data=detected_masks)
detected_boxlist.add_field('scores', detected_scores)
#print('detected_boxlist.num_boxes()',detected_boxlist.num_boxes())
detected_boxlist = np_box_mask_list_ops.non_max_suppression(
detected_boxlist, self.nms_max_output_boxes,
self.nms_iou_threshold)
#print('sfter nms',detected_boxlist.num_boxes())
gt_non_group_of_boxlist = np_box_mask_list.BoxMaskList(
box_data=groundtruth_boxes[~groundtruth_is_group_of_list],
mask_data=groundtruth_masks[~groundtruth_is_group_of_list])
gt_group_of_boxlist = np_box_mask_list.BoxMaskList(
box_data=groundtruth_boxes[groundtruth_is_group_of_list],
mask_data=groundtruth_masks[groundtruth_is_group_of_list])
iou = np_box_mask_list_ops.iou(detected_boxlist,
gt_non_group_of_boxlist)
ioa = np.transpose(
np_box_mask_list_ops.ioa(gt_group_of_boxlist, detected_boxlist))
scores = detected_boxlist.get_field('scores')
num_boxes = detected_boxlist.num_boxes()
return iou, ioa, scores, num_boxes
def test_nms_disabled_max_output_size_equals_one(self):
box_mask_list = np_box_mask_list.BoxMaskList(box_data=self.boxes2,
mask_data=self.masks2)
box_mask_list.add_field('scores', np.array([.9, .75, .6], dtype=float))
max_output_size = 1
iou_threshold = 1. # No NMS
expected_boxes = np.array([[3.0, 4.0, 6.0, 8.0]], dtype=float)
expected_masks = np.array([[[0, 1, 0], [1, 1, 1], [0, 0, 0]]],
dtype=np.uint8)
nms_box_mask_list = np_box_mask_list_ops.non_max_suppression(
box_mask_list, max_output_size, iou_threshold)
self.assertAllClose(nms_box_mask_list.get(), expected_boxes)
self.assertAllClose(nms_box_mask_list.get_masks(), expected_masks)
def test_nms_disabled_max_output_size_equals_one(self):
box_mask_list = np_box_mask_list.BoxMaskList(
box_data=self.boxes2, mask_data=self.masks2)
box_mask_list.add_field('scores',
np.array([.9, .75, .6], dtype=float))
max_output_size = 1
iou_threshold = 1. # No NMS
expected_boxes = np.array([[3.0, 4.0, 6.0, 8.0]], dtype=float)
expected_masks = np.array(
[[[0, 1, 0], [1, 1, 1], [0, 0, 0]]], dtype=np.uint8)
nms_box_mask_list = np_box_mask_list_ops.non_max_suppression(
box_mask_list, max_output_size, iou_threshold)
self.assertAllClose(nms_box_mask_list.get(), expected_boxes)
self.assertAllClose(nms_box_mask_list.get_masks(), expected_masks)
def _get_overlaps_and_scores_mask_mode(
self, detected_boxes, detected_scores, detected_masks, groundtruth_boxes,
groundtruth_masks, groundtruth_is_group_of_list):
"""Computes overlaps and scores between detected and groudntruth masks.
Args:
detected_boxes: A numpy array of shape [N, 4] representing detected box
coordinates
detected_scores: A 1-d numpy array of length N representing classification
score
detected_masks: A uint8 numpy array of shape [N, height, width]. If not
None, the scores will be computed based on masks.
groundtruth_boxes: A numpy array of shape [M, 4] representing ground truth
box coordinates
groundtruth_masks: A uint8 numpy array of shape [M, height, width].
groundtruth_is_group_of_list: A boolean numpy array of length M denoting
whether a ground truth box has group-of tag. If a groundtruth box
is group-of box, every detection matching this box is ignored.
Returns:
iou: A float numpy array of size [num_detected_boxes, num_gt_boxes]. If
gt_non_group_of_boxlist.num_boxes() == 0 it will be None.
ioa: A float numpy array of size [num_detected_boxes, num_gt_boxes]. If
gt_group_of_boxlist.num_boxes() == 0 it will be None.
scores: The score of the detected boxlist.
num_boxes: Number of non-maximum suppressed detected boxes.
"""
detected_boxlist = np_box_mask_list.BoxMaskList(
box_data=detected_boxes, mask_data=detected_masks)
detected_boxlist.add_field('scores', detected_scores)
detected_boxlist = np_box_mask_list_ops.non_max_suppression(
detected_boxlist, self.nms_max_output_boxes, self.nms_iou_threshold)
gt_non_group_of_boxlist = np_box_mask_list.BoxMaskList(
box_data=groundtruth_boxes[~groundtruth_is_group_of_list],
mask_data=groundtruth_masks[~groundtruth_is_group_of_list])
gt_group_of_boxlist = np_box_mask_list.BoxMaskList(
box_data=groundtruth_boxes[groundtruth_is_group_of_list],
mask_data=groundtruth_masks[groundtruth_is_group_of_list])
iou = np_box_mask_list_ops.iou(detected_boxlist, gt_non_group_of_boxlist)
ioa = np.transpose(
np_box_mask_list_ops.ioa(gt_group_of_boxlist, detected_boxlist))
scores = detected_boxlist.get_field('scores')
num_boxes = detected_boxlist.num_boxes()
return iou, ioa, scores, num_boxes
