def test_compute_semantic_labels(self):
inputs = {}
points = tf.constant([[0.0, 0.0, 0.0],
[1.0, 1.0, 1.0],
[1.1, 1.0, 1.0],
[1.0, 1.1, 1.1],
[3.0, 3.0, 3.0],
[3.1, 3.0, 3.0],
[3.0, 3.0, 3.1],
[5.0, 5.0, 5.0]], dtype=tf.float32)
inputs['points'] = points
inputs['objects/shape/dimension'] = tf.constant([[0.5, 0.5, 0.5],
[0.5, 0.5, 0.5],
[0.5, 0.5, 0.5]],
dtype=tf.float32)
inputs['objects/pose/R'] = tf.constant([[[1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0]],
[[1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0]],
[[1.0, 0.0, 0.0],
[0.0, 1.0, 0.0],
[0.0, 0.0, 1.0]]],
dtype=tf.float32)
inputs['objects/pose/t'] = tf.constant([[0.0, 0.0, 0.0],
[1.0, 1.0, 1.0],
[3.0, 3.0, 3.0]], dtype=tf.float32)
inputs['objects/category/label'] = tf.constant([1, 2, 3], dtype=tf.int32)
point_labels = waymo_object_per_frame.compute_semantic_labels(
inputs=inputs, points_key='points', box_margin=0.1)
expected_point_labels = tf.constant(
[[1], [2], [2], [2], [3], [3], [3], [0]], dtype=tf.int32)
self.assertAllEqual(point_labels.numpy(), expected_point_labels.numpy())
def test_compute_semantic_labels_value_error(self):
num_points = 100
num_objects = 10
points = tf.random.uniform([num_points, 3],
minval=-5.0,
maxval=5.0,
dtype=tf.float32)
objects_dimension = tf.random.uniform([num_objects, 3],
minval=0.01,
maxval=10.0,
dtype=tf.float32)
objects_rotation = tf.random.uniform([num_objects, 3, 3],
minval=-1.0,
maxval=1.0,
dtype=tf.float32)
objects_center = tf.random.uniform([num_objects, 3],
minval=1 - .0,
maxval=10.0,
dtype=tf.float32)
objects_label = tf.random.uniform([num_objects],
minval=0,
maxval=10,
dtype=tf.int32)
inputs = {
'objects/shape/dimension': objects_dimension,
'objects/pose/R': objects_rotation,
'objects/pose/t': objects_center,
'objects/category/label': objects_label,
}
with self.assertRaises(ValueError):
waymo_object_per_frame.compute_semantic_labels(
inputs=inputs, points_key='points', box_margin=0.1)
inputs = {
'points': points,
'objects/pose/R': objects_rotation,
'objects/pose/t': objects_center,
'objects/category/label': objects_label,
}
with self.assertRaises(ValueError):
waymo_object_per_frame.compute_semantic_labels(
inputs=inputs, points_key='points', box_margin=0.1)
inputs = {
'points': points,
'objects/shape/dimension': objects_dimension,
'objects/pose/t': objects_center,
'objects/category/label': objects_label,
}
with self.assertRaises(ValueError):
waymo_object_per_frame.compute_semantic_labels(
inputs=inputs, points_key='points', box_margin=0.1)
inputs = {
'points': points,
'objects/shape/dimension': objects_dimension,
'objects/pose/R': objects_rotation,
'objects/category/label': objects_label,
}
with self.assertRaises(ValueError):
waymo_object_per_frame.compute_semantic_labels(
inputs=inputs, points_key='points', box_margin=0.1)
inputs = {
'points': points,
'objects/shape/dimension': objects_dimension,
'objects/pose/R': objects_rotation,
'objects/pose/t': objects_center,
}
with self.assertRaises(ValueError):
waymo_object_per_frame.compute_semantic_labels(
inputs=inputs, points_key='points', box_margin=0.1)