Python object_detection.core.preprocessor.resize_image() Examples

def testResizeImageWithMasks(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[15, 50, 100], [10, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks, _ = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def testResizeImageWithNoInstanceMask(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[0, 60, 40], [0, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[0, 50, 100], [0, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def test_build_resize_image(self):
    preprocessor_text_proto = """
    resize_image {
      new_height: 75
      new_width: 100
      method: BICUBIC
    }
    """
    preprocessor_proto = preprocessor_pb2.PreprocessingStep()
    text_format.Merge(preprocessor_text_proto, preprocessor_proto)
    function, args = preprocessor_builder.build(preprocessor_proto)
    self.assertEqual(function, preprocessor.resize_image)
    self.assertEqual(args, {'new_height': 75,
                            'new_width': 100,
                            'method': tf.image.ResizeMethod.BICUBIC}) 

def test_build_resize_image(self):
    preprocessor_text_proto = """
    resize_image {
      new_height: 75
      new_width: 100
      method: BICUBIC
    }
    """
    preprocessor_proto = preprocessor_pb2.PreprocessingStep()
    text_format.Merge(preprocessor_text_proto, preprocessor_proto)
    function, args = preprocessor_builder.build(preprocessor_proto)
    self.assertEqual(function, preprocessor.resize_image)
    self.assertEqual(args, {'new_height': 75,
                            'new_width': 100,
                            'method': tf.image.ResizeMethod.BICUBIC}) 

def test_build_resize_image(self):
    preprocessor_text_proto = """
    resize_image {
      new_height: 75
      new_width: 100
      method: BICUBIC
    }
    """
    preprocessor_proto = preprocessor_pb2.PreprocessingStep()
    text_format.Merge(preprocessor_text_proto, preprocessor_proto)
    function, args = preprocessor_builder.build(preprocessor_proto)
    self.assertEqual(function, preprocessor.resize_image)
    self.assertEqual(args, {'new_height': 75,
                            'new_width': 100,
                            'method': tf.image.ResizeMethod.BICUBIC}) 

def build(image_resizer_config):
  """Builds callable for image resizing operations.

  Args:
    image_resizer_config: image_resizer.proto object containing parameters for
      an image resizing operation.

  Returns:
    image_resizer_fn: Callable for image resizing.  This callable always takes
      a rank-3 image tensor (corresponding to a single image) and returns a
      rank-3 image tensor, possibly with new spatial dimensions.

  Raises:
    ValueError: if `image_resizer_config` is of incorrect type.
    ValueError: if `image_resizer_config.image_resizer_oneof` is of expected
      type.
    ValueError: if min_dimension > max_dimension when keep_aspect_ratio_resizer
      is used.
  """
  if not isinstance(image_resizer_config, image_resizer_pb2.ImageResizer):
    raise ValueError('image_resizer_config not of type '
                     'image_resizer_pb2.ImageResizer.')

  if image_resizer_config.WhichOneof(
      'image_resizer_oneof') == 'keep_aspect_ratio_resizer':
    keep_aspect_ratio_config = image_resizer_config.keep_aspect_ratio_resizer
    if not (keep_aspect_ratio_config.min_dimension
            <= keep_aspect_ratio_config.max_dimension):
      raise ValueError('min_dimension > max_dimension')
    return functools.partial(
        preprocessor.resize_to_range,
        min_dimension=keep_aspect_ratio_config.min_dimension,
        max_dimension=keep_aspect_ratio_config.max_dimension)
  if image_resizer_config.WhichOneof(
      'image_resizer_oneof') == 'fixed_shape_resizer':
    fixed_shape_resizer_config = image_resizer_config.fixed_shape_resizer
    return functools.partial(preprocessor.resize_image,
                             new_height=fixed_shape_resizer_config.height,
                             new_width=fixed_shape_resizer_config.width)
  raise ValueError('Invalid image resizer option.') 

def test_build_resize_image(self):
    preprocessor_text_proto = """
    resize_image {
      new_height: 75
      new_width: 100
      method: BICUBIC
    }
    """
    preprocessor_proto = preprocessor_pb2.PreprocessingStep()
    text_format.Merge(preprocessor_text_proto, preprocessor_proto)
    function, args = preprocessor_builder.build(preprocessor_proto)
    self.assertEqual(function, preprocessor.resize_image)
    self.assertEqual(args, {'new_height': 75,
                            'new_width': 100,
                            'method': tf.image.ResizeMethod.BICUBIC}) 

def build(image_resizer_config):
  """Builds callable for image resizing operations.

  Args:
    image_resizer_config: image_resizer.proto object containing parameters for
      an image resizing operation.

  Returns:
    image_resizer_fn: Callable for image resizing.  This callable always takes
      a rank-3 image tensor (corresponding to a single image) and returns a
      rank-3 image tensor, possibly with new spatial dimensions.

  Raises:
    ValueError: if `image_resizer_config` is of incorrect type.
    ValueError: if `image_resizer_config.image_resizer_oneof` is of expected
      type.
    ValueError: if min_dimension > max_dimension when keep_aspect_ratio_resizer
      is used.
  """
  if not isinstance(image_resizer_config, image_resizer_pb2.ImageResizer):
    raise ValueError('image_resizer_config not of type '
                     'image_resizer_pb2.ImageResizer.')

  if image_resizer_config.WhichOneof(
      'image_resizer_oneof') == 'keep_aspect_ratio_resizer':
    keep_aspect_ratio_config = image_resizer_config.keep_aspect_ratio_resizer
    if not (keep_aspect_ratio_config.min_dimension
            <= keep_aspect_ratio_config.max_dimension):
      raise ValueError('min_dimension > max_dimension')
    return functools.partial(
        preprocessor.resize_to_range,
        min_dimension=keep_aspect_ratio_config.min_dimension,
        max_dimension=keep_aspect_ratio_config.max_dimension)
  if image_resizer_config.WhichOneof(
      'image_resizer_oneof') == 'fixed_shape_resizer':
    fixed_shape_resizer_config = image_resizer_config.fixed_shape_resizer
    return functools.partial(preprocessor.resize_image,
                             new_height=fixed_shape_resizer_config.height,
                             new_width=fixed_shape_resizer_config.width)
  raise ValueError('Invalid image resizer option.') 

def testResizeImageWithNoInstanceMask(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[0, 60, 40], [0, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[0, 50, 100], [0, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def testResizeImageWithMasks(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[15, 50, 100], [10, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def testResizeImageWithMasks(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[15, 50, 100], [10, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def test_include_masks_in_data_augmentation(self):
    data_augmentation_options = [
        (preprocessor.resize_image, {
            'new_height': 20,
            'new_width': 20,
            'method': tf.image.ResizeMethod.NEAREST_NEIGHBOR
        })
    ]
    data_augmentation_fn = functools.partial(
        inputs.augment_input_data,
        data_augmentation_options=data_augmentation_options)
    tensor_dict = {
        fields.InputDataFields.image:
            tf.constant(np.random.rand(10, 10, 3).astype(np.float32)),
        fields.InputDataFields.groundtruth_instance_masks:
            tf.constant(np.zeros([2, 10, 10], np.uint8))
    }
    augmented_tensor_dict = data_augmentation_fn(tensor_dict=tensor_dict)
    with self.test_session() as sess:
      augmented_tensor_dict_out = sess.run(augmented_tensor_dict)

    self.assertAllEqual(
        augmented_tensor_dict_out[fields.InputDataFields.image].shape,
        [20, 20, 3])
    self.assertAllEqual(augmented_tensor_dict_out[
        fields.InputDataFields.groundtruth_instance_masks].shape, [2, 20, 20]) 

def test_apply_image_and_box_augmentation(self):
    data_augmentation_options = [
        (preprocessor.resize_image, {
            'new_height': 20,
            'new_width': 20,
            'method': tf.image.ResizeMethod.NEAREST_NEIGHBOR
        }),
        (preprocessor.scale_boxes_to_pixel_coordinates, {}),
    ]
    data_augmentation_fn = functools.partial(
        inputs.augment_input_data,
        data_augmentation_options=data_augmentation_options)
    tensor_dict = {
        fields.InputDataFields.image:
            tf.constant(np.random.rand(10, 10, 3).astype(np.float32)),
        fields.InputDataFields.groundtruth_boxes:
            tf.constant(np.array([[.5, .5, 1., 1.]], np.float32))
    }
    augmented_tensor_dict = data_augmentation_fn(tensor_dict=tensor_dict)
    with self.test_session() as sess:
      augmented_tensor_dict_out = sess.run(augmented_tensor_dict)

    self.assertAllEqual(
        augmented_tensor_dict_out[fields.InputDataFields.image].shape,
        [20, 20, 3]
    )
    self.assertAllClose(
        augmented_tensor_dict_out[fields.InputDataFields.groundtruth_boxes],
        [[10, 10, 20, 20]]
    ) 

def test_build_resize_image(self):
    preprocessor_text_proto = """
    resize_image {
      new_height: 75
      new_width: 100
      method: BICUBIC
    }
    """
    preprocessor_proto = preprocessor_pb2.PreprocessingStep()
    text_format.Merge(preprocessor_text_proto, preprocessor_proto)
    function, args = preprocessor_builder.build(preprocessor_proto)
    self.assertEqual(function, preprocessor.resize_image)
    self.assertEqual(args, {'new_height': 75,
                            'new_width': 100,
                            'method': tf.image.ResizeMethod.BICUBIC}) 

def testResizeImageWithMasksTensorInputHeightAndWidth(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
    height = tf.constant(50, dtype=tf.int32)
    width = tf.constant(100, dtype=tf.int32)
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[15, 50, 100], [10, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks, _ = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def test_include_keypoints_in_data_augmentation(self):
    data_augmentation_options = [
        (preprocessor.resize_image, {
            'new_height': 20,
            'new_width': 20,
            'method': tf.image.ResizeMethod.NEAREST_NEIGHBOR
        }),
        (preprocessor.scale_boxes_to_pixel_coordinates, {}),
    ]
    data_augmentation_fn = functools.partial(
        inputs.augment_input_data,
        data_augmentation_options=data_augmentation_options)
    tensor_dict = {
        fields.InputDataFields.image:
            tf.constant(np.random.rand(10, 10, 3).astype(np.float32)),
        fields.InputDataFields.groundtruth_boxes:
            tf.constant(np.array([[.5, .5, 1., 1.]], np.float32)),
        fields.InputDataFields.groundtruth_keypoints:
            tf.constant(np.array([[[0.5, 1.0], [0.5, 0.5]]], np.float32))
    }
    augmented_tensor_dict = data_augmentation_fn(tensor_dict=tensor_dict)
    with self.test_session() as sess:
      augmented_tensor_dict_out = sess.run(augmented_tensor_dict)

    self.assertAllEqual(
        augmented_tensor_dict_out[fields.InputDataFields.image].shape,
        [20, 20, 3]
    )
    self.assertAllClose(
        augmented_tensor_dict_out[fields.InputDataFields.groundtruth_boxes],
        [[10, 10, 20, 20]]
    )
    self.assertAllClose(
        augmented_tensor_dict_out[fields.InputDataFields.groundtruth_keypoints],
        [[[10, 20], [10, 10]]]
    ) 

def test_include_keypoints_in_data_augmentation(self):
    data_augmentation_options = [
        (preprocessor.resize_image, {
            'new_height': 20,
            'new_width': 20,
            'method': tf.image.ResizeMethod.NEAREST_NEIGHBOR
        }),
        (preprocessor.scale_boxes_to_pixel_coordinates, {}),
    ]
    data_augmentation_fn = functools.partial(
        inputs.augment_input_data,
        data_augmentation_options=data_augmentation_options)
    tensor_dict = {
        fields.InputDataFields.image:
            tf.constant(np.random.rand(10, 10, 3).astype(np.float32)),
        fields.InputDataFields.groundtruth_boxes:
            tf.constant(np.array([[.5, .5, 1., 1.]], np.float32)),
        fields.InputDataFields.groundtruth_keypoints:
            tf.constant(np.array([[[0.5, 1.0], [0.5, 0.5]]], np.float32))
    }
    augmented_tensor_dict = data_augmentation_fn(tensor_dict=tensor_dict)
    with self.test_session() as sess:
      augmented_tensor_dict_out = sess.run(augmented_tensor_dict)

    self.assertAllEqual(
        augmented_tensor_dict_out[fields.InputDataFields.image].shape,
        [20, 20, 3]
    )
    self.assertAllClose(
        augmented_tensor_dict_out[fields.InputDataFields.groundtruth_boxes],
        [[10, 10, 20, 20]]
    )
    self.assertAllClose(
        augmented_tensor_dict_out[fields.InputDataFields.groundtruth_keypoints],
        [[[10, 20], [10, 10]]]
    ) 

def test_include_masks_in_data_augmentation(self):
    data_augmentation_options = [
        (preprocessor.resize_image, {
            'new_height': 20,
            'new_width': 20,
            'method': tf.image.ResizeMethod.NEAREST_NEIGHBOR
        })
    ]
    data_augmentation_fn = functools.partial(
        inputs.augment_input_data,
        data_augmentation_options=data_augmentation_options)
    tensor_dict = {
        fields.InputDataFields.image:
            tf.constant(np.random.rand(10, 10, 3).astype(np.float32)),
        fields.InputDataFields.groundtruth_instance_masks:
            tf.constant(np.zeros([2, 10, 10], np.uint8))
    }
    augmented_tensor_dict = data_augmentation_fn(tensor_dict=tensor_dict)
    with self.test_session() as sess:
      augmented_tensor_dict_out = sess.run(augmented_tensor_dict)

    self.assertAllEqual(
        augmented_tensor_dict_out[fields.InputDataFields.image].shape,
        [20, 20, 3])
    self.assertAllEqual(augmented_tensor_dict_out[
        fields.InputDataFields.groundtruth_instance_masks].shape, [2, 20, 20]) 

def test_apply_image_and_box_augmentation(self):
    data_augmentation_options = [
        (preprocessor.resize_image, {
            'new_height': 20,
            'new_width': 20,
            'method': tf.image.ResizeMethod.NEAREST_NEIGHBOR
        }),
        (preprocessor.scale_boxes_to_pixel_coordinates, {}),
    ]
    data_augmentation_fn = functools.partial(
        inputs.augment_input_data,
        data_augmentation_options=data_augmentation_options)
    tensor_dict = {
        fields.InputDataFields.image:
            tf.constant(np.random.rand(10, 10, 3).astype(np.float32)),
        fields.InputDataFields.groundtruth_boxes:
            tf.constant(np.array([[.5, .5, 1., 1.]], np.float32))
    }
    augmented_tensor_dict = data_augmentation_fn(tensor_dict=tensor_dict)
    with self.test_session() as sess:
      augmented_tensor_dict_out = sess.run(augmented_tensor_dict)

    self.assertAllEqual(
        augmented_tensor_dict_out[fields.InputDataFields.image].shape,
        [20, 20, 3]
    )
    self.assertAllClose(
        augmented_tensor_dict_out[fields.InputDataFields.groundtruth_boxes],
        [[10, 10, 20, 20]]
    ) 

def test_build_resize_image(self):
    preprocessor_text_proto = """
    resize_image {
      new_height: 75
      new_width: 100
      method: BICUBIC
    }
    """
    preprocessor_proto = preprocessor_pb2.PreprocessingStep()
    text_format.Merge(preprocessor_text_proto, preprocessor_proto)
    function, args = preprocessor_builder.build(preprocessor_proto)
    self.assertEqual(function, preprocessor.resize_image)
    self.assertEqual(args, {'new_height': 75,
                            'new_width': 100,
                            'method': tf.image.ResizeMethod.BICUBIC}) 

def build(image_resizer_config):
  """Builds callable for image resizing operations.

  Args:
    image_resizer_config: image_resizer.proto object containing parameters for
      an image resizing operation.

  Returns:
    image_resizer_fn: Callable for image resizing.  This callable always takes
      a rank-3 image tensor (corresponding to a single image) and returns a
      rank-3 image tensor, possibly with new spatial dimensions.

  Raises:
    ValueError: if `image_resizer_config` is of incorrect type.
    ValueError: if `image_resizer_config.image_resizer_oneof` is of expected
      type.
    ValueError: if min_dimension > max_dimension when keep_aspect_ratio_resizer
      is used.
  """
  if not isinstance(image_resizer_config, image_resizer_pb2.ImageResizer):
    raise ValueError('image_resizer_config not of type '
                     'image_resizer_pb2.ImageResizer.')

  if image_resizer_config.WhichOneof(
      'image_resizer_oneof') == 'keep_aspect_ratio_resizer':
    keep_aspect_ratio_config = image_resizer_config.keep_aspect_ratio_resizer
    if not (keep_aspect_ratio_config.min_dimension
            <= keep_aspect_ratio_config.max_dimension):
      raise ValueError('min_dimension > max_dimension')
    return functools.partial(
        preprocessor.resize_to_range,
        min_dimension=keep_aspect_ratio_config.min_dimension,
        max_dimension=keep_aspect_ratio_config.max_dimension)
  if image_resizer_config.WhichOneof(
      'image_resizer_oneof') == 'fixed_shape_resizer':
    fixed_shape_resizer_config = image_resizer_config.fixed_shape_resizer
    return functools.partial(preprocessor.resize_image,
                             new_height=fixed_shape_resizer_config.height,
                             new_width=fixed_shape_resizer_config.width)
  raise ValueError('Invalid image resizer option.') 

def testResizeToRangeWithInstanceMasksTensorOfSizeZero(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[0, 60, 40], [0, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[0, 50, 100], [0, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def test_build_resize_image(self):
    preprocessor_text_proto = """
    resize_image {
      new_height: 75
      new_width: 100
      method: BICUBIC
    }
    """
    preprocessor_proto = preprocessor_pb2.PreprocessingStep()
    text_format.Merge(preprocessor_text_proto, preprocessor_proto)
    function, args = preprocessor_builder.build(preprocessor_proto)
    self.assertEqual(function, preprocessor.resize_image)
    self.assertEqual(args, {'new_height': 75,
                            'new_width': 100,
                            'method': tf.image.ResizeMethod.BICUBIC}) 

def build(image_resizer_config):
  """Builds callable for image resizing operations.

  Args:
    image_resizer_config: image_resizer.proto object containing parameters for
      an image resizing operation.

  Returns:
    image_resizer_fn: Callable for image resizing.  This callable always takes
      a rank-3 image tensor (corresponding to a single image) and returns a
      rank-3 image tensor, possibly with new spatial dimensions.

  Raises:
    ValueError: if `image_resizer_config` is of incorrect type.
    ValueError: if `image_resizer_config.image_resizer_oneof` is of expected
      type.
    ValueError: if min_dimension > max_dimension when keep_aspect_ratio_resizer
      is used.
  """
  if not isinstance(image_resizer_config, image_resizer_pb2.ImageResizer):
    raise ValueError('image_resizer_config not of type '
                     'image_resizer_pb2.ImageResizer.')

  if image_resizer_config.WhichOneof(
      'image_resizer_oneof') == 'keep_aspect_ratio_resizer':
    keep_aspect_ratio_config = image_resizer_config.keep_aspect_ratio_resizer
    if not (keep_aspect_ratio_config.min_dimension
            <= keep_aspect_ratio_config.max_dimension):
      raise ValueError('min_dimension > max_dimension')
    return functools.partial(
        preprocessor.resize_to_range,
        min_dimension=keep_aspect_ratio_config.min_dimension,
        max_dimension=keep_aspect_ratio_config.max_dimension)
  if image_resizer_config.WhichOneof(
      'image_resizer_oneof') == 'fixed_shape_resizer':
    fixed_shape_resizer_config = image_resizer_config.fixed_shape_resizer
    return functools.partial(preprocessor.resize_image,
                             new_height=fixed_shape_resizer_config.height,
                             new_width=fixed_shape_resizer_config.width)
  raise ValueError('Invalid image resizer option.') 

def testResizeToRangeWithInstanceMasksTensorOfSizeZero(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[0, 60, 40], [0, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[0, 50, 100], [0, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def test_build_resize_image(self):
    preprocessor_text_proto = """
    resize_image {
      new_height: 75
      new_width: 100
      method: BICUBIC
    }
    """
    preprocessor_proto = preprocessor_pb2.PreprocessingStep()
    text_format.Merge(preprocessor_text_proto, preprocessor_proto)
    function, args = preprocessor_builder.build(preprocessor_proto)
    self.assertEqual(function, preprocessor.resize_image)
    self.assertEqual(args, {'new_height': 75,
                            'new_width': 100,
                            'method': tf.image.ResizeMethod.BICUBIC}) 

def testResizeImageWithNoInstanceMask(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[0, 60, 40], [0, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[0, 50, 100], [0, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks, _ = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def testResizeImageWithMasks(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[15, 50, 100], [10, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks, _ = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape) 

def test_include_keypoints_in_data_augmentation(self):
    data_augmentation_options = [
        (preprocessor.resize_image, {
            'new_height': 20,
            'new_width': 20,
            'method': tf.image.ResizeMethod.NEAREST_NEIGHBOR
        }),
        (preprocessor.scale_boxes_to_pixel_coordinates, {}),
    ]
    data_augmentation_fn = functools.partial(
        inputs.augment_input_data,
        data_augmentation_options=data_augmentation_options)
    tensor_dict = {
        fields.InputDataFields.image:
            tf.constant(np.random.rand(10, 10, 3).astype(np.float32)),
        fields.InputDataFields.groundtruth_boxes:
            tf.constant(np.array([[.5, .5, 1., 1.]], np.float32)),
        fields.InputDataFields.groundtruth_keypoints:
            tf.constant(np.array([[[0.5, 1.0], [0.5, 0.5]]], np.float32))
    }
    augmented_tensor_dict = data_augmentation_fn(tensor_dict=tensor_dict)
    with self.test_session() as sess:
      augmented_tensor_dict_out = sess.run(augmented_tensor_dict)

    self.assertAllEqual(
        augmented_tensor_dict_out[fields.InputDataFields.image].shape,
        [20, 20, 3]
    )
    self.assertAllClose(
        augmented_tensor_dict_out[fields.InputDataFields.groundtruth_boxes],
        [[10, 10, 20, 20]]
    )
    self.assertAllClose(
        augmented_tensor_dict_out[fields.InputDataFields.groundtruth_keypoints],
        [[[10, 20], [10, 10]]]
    ) 

def testResizeImageWithMasks(self):
    """Tests image resizing, checking output sizes."""
    in_image_shape_list = [[60, 40, 3], [15, 30, 3]]
    in_masks_shape_list = [[15, 60, 40], [10, 15, 30]]
    height = 50
    width = 100
    expected_image_shape_list = [[50, 100, 3], [50, 100, 3]]
    expected_masks_shape_list = [[15, 50, 100], [10, 50, 100]]

    for (in_image_shape, expected_image_shape, in_masks_shape,
         expected_mask_shape) in zip(in_image_shape_list,
                                     expected_image_shape_list,
                                     in_masks_shape_list,
                                     expected_masks_shape_list):
      in_image = tf.random_uniform(in_image_shape)
      in_masks = tf.random_uniform(in_masks_shape)
      out_image, out_masks = preprocessor.resize_image(
          in_image, in_masks, new_height=height, new_width=width)
      out_image_shape = tf.shape(out_image)
      out_masks_shape = tf.shape(out_masks)

      with self.test_session() as sess:
        out_image_shape, out_masks_shape = sess.run(
            [out_image_shape, out_masks_shape])
        self.assertAllEqual(out_image_shape, expected_image_shape)
        self.assertAllEqual(out_masks_shape, expected_mask_shape)