Question

在张量流中，我想从随机角度旋转图像，以进行数据增强。但我在tf.image模块中找不到这种转换。

Answer 1

这可以在tensorflow now中完成：

tf.contrib.image.rotate(images, degrees * math.pi / 180, interpolation='BILINEAR')

Answer 2

因为我想能够旋转张量，所以我提出了以下代码，它将[高度，宽度，深度]张量旋转了一个给定的角度：

def rotate_image_tensor(image, angle, mode='black'):
    """
    Rotates a 3D tensor (HWD), which represents an image by given radian angle.

    New image has the same size as the input image.

    mode controls what happens to border pixels.
    mode = 'black' results in black bars (value 0 in unknown areas)
    mode = 'white' results in value 255 in unknown areas
    mode = 'ones' results in value 1 in unknown areas
    mode = 'repeat' keeps repeating the closest pixel known
    """
    s = image.get_shape().as_list()
    assert len(s) == 3, "Input needs to be 3D."
    assert (mode == 'repeat') or (mode == 'black') or (mode == 'white') or (mode == 'ones'), "Unknown boundary mode."
    image_center = [np.floor(x/2) for x in s]

    # Coordinates of new image
    coord1 = tf.range(s[0])
    coord2 = tf.range(s[1])

    # Create vectors of those coordinates in order to vectorize the image
    coord1_vec = tf.tile(coord1, [s[1]])

    coord2_vec_unordered = tf.tile(coord2, [s[0]])
    coord2_vec_unordered = tf.reshape(coord2_vec_unordered, [s[0], s[1]])
    coord2_vec = tf.reshape(tf.transpose(coord2_vec_unordered, [1, 0]), [-1])

    # center coordinates since rotation center is supposed to be in the image center
    coord1_vec_centered = coord1_vec - image_center[0]
    coord2_vec_centered = coord2_vec - image_center[1]

    coord_new_centered = tf.cast(tf.pack([coord1_vec_centered, coord2_vec_centered]), tf.float32)

    # Perform backward transformation of the image coordinates
    rot_mat_inv = tf.dynamic_stitch([[0], [1], [2], [3]], [tf.cos(angle), tf.sin(angle), -tf.sin(angle), tf.cos(angle)])
    rot_mat_inv = tf.reshape(rot_mat_inv, shape=[2, 2])
    coord_old_centered = tf.matmul(rot_mat_inv, coord_new_centered)

    # Find nearest neighbor in old image
    coord1_old_nn = tf.cast(tf.round(coord_old_centered[0, :] + image_center[0]), tf.int32)
    coord2_old_nn = tf.cast(tf.round(coord_old_centered[1, :] + image_center[1]), tf.int32)

    # Clip values to stay inside image coordinates
    if mode == 'repeat':
        coord_old1_clipped = tf.minimum(tf.maximum(coord1_old_nn, 0), s[0]-1)
        coord_old2_clipped = tf.minimum(tf.maximum(coord2_old_nn, 0), s[1]-1)
    else:
        outside_ind1 = tf.logical_or(tf.greater(coord1_old_nn, s[0]-1), tf.less(coord1_old_nn, 0))
        outside_ind2 = tf.logical_or(tf.greater(coord2_old_nn, s[1]-1), tf.less(coord2_old_nn, 0))
        outside_ind = tf.logical_or(outside_ind1, outside_ind2)

        coord_old1_clipped = tf.boolean_mask(coord1_old_nn, tf.logical_not(outside_ind))
        coord_old2_clipped = tf.boolean_mask(coord2_old_nn, tf.logical_not(outside_ind))

        coord1_vec = tf.boolean_mask(coord1_vec, tf.logical_not(outside_ind))
        coord2_vec = tf.boolean_mask(coord2_vec, tf.logical_not(outside_ind))

    coord_old_clipped = tf.cast(tf.transpose(tf.pack([coord_old1_clipped, coord_old2_clipped]), [1, 0]), tf.int32)

    # Coordinates of the new image
    coord_new = tf.transpose(tf.cast(tf.pack([coord1_vec, coord2_vec]), tf.int32), [1, 0])

    image_channel_list = tf.split(2, s[2], image)

    image_rotated_channel_list = list()
    for image_channel in image_channel_list:
        image_chan_new_values = tf.gather_nd(tf.squeeze(image_channel), coord_old_clipped)

        if (mode == 'black') or (mode == 'repeat'):
            background_color = 0
        elif mode == 'ones':
            background_color = 1
        elif mode == 'white':
            background_color = 255

        image_rotated_channel_list.append(tf.sparse_to_dense(coord_new, [s[0], s[1]], image_chan_new_values,
                                                             background_color, validate_indices=False))

    image_rotated = tf.transpose(tf.pack(image_rotated_channel_list), [1, 2, 0])

    return image_rotated

Answer 3

更新：请参阅下面的@ astromme答案。 Tensorflow现在支持原生旋转图像。

在tensorflow中没有本机方法时你可以做的是这样的：

from PIL import Image
sess = tf.InteractiveSession()

# Pass image tensor object to a PIL image
image = Image.fromarray(image.eval())

# Use PIL or other library of the sort to rotate
rotated = Image.Image.rotate(image, degrees)

# Convert rotated image back to tensor
rotated_tensor = tf.convert_to_tensor(np.array(rotated))

Answer 4

TensorFlow中的旋转和裁剪

我个人需要在TensorFlow中进行图像旋转和裁剪黑色边框功能，如下所示。我可以实现以下功能。

def _rotate_and_crop(image, output_height, output_width, rotation_degree, do_crop):
    """Rotate the given image with the given rotation degree and crop for the black edges if necessary
    Args:
        image: A `Tensor` representing an image of arbitrary size.
        output_height: The height of the image after preprocessing.
        output_width: The width of the image after preprocessing.
        rotation_degree: The degree of rotation on the image.
        do_crop: Do cropping if it is True.
    Returns:
        A rotated image.
    """

    # Rotate the given image with the given rotation degree
    if rotation_degree != 0:
        image = tf.contrib.image.rotate(image, math.radians(rotation_degree), interpolation='BILINEAR')

        # Center crop to ommit black noise on the edges
        if do_crop == True:
            lrr_width, lrr_height = _largest_rotated_rect(output_height, output_width, math.radians(rotation_degree))
            resized_image = tf.image.central_crop(image, float(lrr_height)/output_height)    
            image = tf.image.resize_images(resized_image, [output_height, output_width], method=tf.image.ResizeMethod.BILINEAR, align_corners=False)

    return image

def _largest_rotated_rect(w, h, angle):
    """
    Given a rectangle of size wxh that has been rotated by 'angle' (in
    radians), computes the width and height of the largest possible
    axis-aligned rectangle within the rotated rectangle.
    Original JS code by 'Andri' and Magnus Hoff from Stack Overflow
    Converted to Python by Aaron Snoswell
    Source: http://stackoverflow.com/questions/16702966/rotate-image-and-crop-out-black-borders
    """

    quadrant = int(math.floor(angle / (math.pi / 2))) & 3
    sign_alpha = angle if ((quadrant & 1) == 0) else math.pi - angle
    alpha = (sign_alpha % math.pi + math.pi) % math.pi

    bb_w = w * math.cos(alpha) + h * math.sin(alpha)
    bb_h = w * math.sin(alpha) + h * math.cos(alpha)

    gamma = math.atan2(bb_w, bb_w) if (w < h) else math.atan2(bb_w, bb_w)

    delta = math.pi - alpha - gamma

    length = h if (w < h) else w

    d = length * math.cos(alpha)
    a = d * math.sin(alpha) / math.sin(delta)

    y = a * math.cos(gamma)
    x = y * math.tan(gamma)

    return (
        bb_w - 2 * x,
        bb_h - 2 * y
    )

如果您需要在TensorFlow中进一步实现示例和可视化，可以使用this repository。我希望这可以对其他人有所帮助。

Answer 5

对于tensorflow 2.0：

Command failed: npm install --save-dev --loglevel error ember-cli-mirage
npm WARN deprecated core-js@2.4.1: core-js@<2.6.8 is no longer maintained. Please, upgrade to core-js@3 or at least to actual version of core-js@2.
npm ERR! Maximum call stack size exceeded

npm ERR! A complete log of this run can be found in:
npm ERR!     /Users/vitaliy/.npm/_logs/2019-08-13T10_31_36_432Z-debug.log

Answer 6

以下是更新为Tensorflow v0.12的@zimmermc答案

的变化：

pack()现在是stack()

split参数的顺序颠倒

def rotate_image_tensor(image, angle, mode='white'):
    """
    Rotates a 3D tensor (HWD), which represents an image by given radian angle.

    New image has the same size as the input image.

    mode controls what happens to border pixels.
    mode = 'black' results in black bars (value 0 in unknown areas)
    mode = 'white' results in value 255 in unknown areas
    mode = 'ones' results in value 1 in unknown areas
    mode = 'repeat' keeps repeating the closest pixel known
    """
    s = image.get_shape().as_list()
    assert len(s) == 3, "Input needs to be 3D."
    assert (mode == 'repeat') or (mode == 'black') or (mode == 'white') or (mode == 'ones'), "Unknown boundary mode."
    image_center = [np.floor(x/2) for x in s]

    # Coordinates of new image
    coord1 = tf.range(s[0])
    coord2 = tf.range(s[1])

    # Create vectors of those coordinates in order to vectorize the image
    coord1_vec = tf.tile(coord1, [s[1]])

    coord2_vec_unordered = tf.tile(coord2, [s[0]])
    coord2_vec_unordered = tf.reshape(coord2_vec_unordered, [s[0], s[1]])
    coord2_vec = tf.reshape(tf.transpose(coord2_vec_unordered, [1, 0]), [-1])

    # center coordinates since rotation center is supposed to be in the image center
    coord1_vec_centered = coord1_vec - image_center[0]
    coord2_vec_centered = coord2_vec - image_center[1]

    coord_new_centered = tf.cast(tf.stack([coord1_vec_centered, coord2_vec_centered]), tf.float32)

    # Perform backward transformation of the image coordinates
    rot_mat_inv = tf.dynamic_stitch([[0], [1], [2], [3]], [tf.cos(angle), tf.sin(angle), -tf.sin(angle), tf.cos(angle)])
    rot_mat_inv = tf.reshape(rot_mat_inv, shape=[2, 2])
    coord_old_centered = tf.matmul(rot_mat_inv, coord_new_centered)

    # Find nearest neighbor in old image
    coord1_old_nn = tf.cast(tf.round(coord_old_centered[0, :] + image_center[0]), tf.int32)
    coord2_old_nn = tf.cast(tf.round(coord_old_centered[1, :] + image_center[1]), tf.int32)

    # Clip values to stay inside image coordinates
    if mode == 'repeat':
        coord_old1_clipped = tf.minimum(tf.maximum(coord1_old_nn, 0), s[0]-1)
        coord_old2_clipped = tf.minimum(tf.maximum(coord2_old_nn, 0), s[1]-1)
    else:
        outside_ind1 = tf.logical_or(tf.greater(coord1_old_nn, s[0]-1), tf.less(coord1_old_nn, 0))
        outside_ind2 = tf.logical_or(tf.greater(coord2_old_nn, s[1]-1), tf.less(coord2_old_nn, 0))
        outside_ind = tf.logical_or(outside_ind1, outside_ind2)

        coord_old1_clipped = tf.boolean_mask(coord1_old_nn, tf.logical_not(outside_ind))
        coord_old2_clipped = tf.boolean_mask(coord2_old_nn, tf.logical_not(outside_ind))

        coord1_vec = tf.boolean_mask(coord1_vec, tf.logical_not(outside_ind))
        coord2_vec = tf.boolean_mask(coord2_vec, tf.logical_not(outside_ind))

    coord_old_clipped = tf.cast(tf.transpose(tf.stack([coord_old1_clipped, coord_old2_clipped]), [1, 0]), tf.int32)

    # Coordinates of the new image
    coord_new = tf.transpose(tf.cast(tf.stack([coord1_vec, coord2_vec]), tf.int32), [1, 0])

    image_channel_list = tf.split(image, s[2], 2)

    image_rotated_channel_list = list()
    for image_channel in image_channel_list:
        image_chan_new_values = tf.gather_nd(tf.squeeze(image_channel), coord_old_clipped)

        if (mode == 'black') or (mode == 'repeat'):
            background_color = 0
        elif mode == 'ones':
            background_color = 1
        elif mode == 'white':
            background_color = 255

        image_rotated_channel_list.append(tf.sparse_to_dense(coord_new, [s[0], s[1]], image_chan_new_values,
                                                             background_color, validate_indices=False))

    image_rotated = tf.transpose(tf.stack(image_rotated_channel_list), [1, 2, 0])

    return image_rotated

Answer 7

要将图像或一批图像逆时针旋转90度的倍数，可以使用tf.image.rot90(image,k=1,name=None)。

k表示您要旋转90度。

对于单个图像，image是3-D Tensor of shape [height, width, channels]，对于一批图像，image是4-D Tensor of shape [batch, height, width, channels]

Answer 8

tf.contrib 在 tensorflow 2 中不可用。

对于 tensorflow >= 2.* 可以使用以下内容：

tf.keras.preprocessing.image.random_rotation(x, rg, row_axis=1,col_axis=2, channel_axis=0,fill_mode='nearest', cval=0., interpolation_order=1);

您可以在这里找到文档： https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/image/random_rotation

tensorflow：如何旋转图像进行数据扩充？

8 个答案:

TensorFlow中的旋转和裁剪