R/transforms-generics.R

In torchvision: Models, Datasets and Transformations for Images

#' Convert an image to a tensor
#'
#' Converts a Magick Image or array (H x W x C) in the range `[0, 255]` to a
#'   `torch_tensor` of shape (C x H x W) in the range `[0.0, 1.0]`. In the
#'   other cases, tensors are returned without scaling.
#'
#' @note
#' Because the input image is scaled to `[0.0, 1.0]`, this transformation
#'   should not be used when transforming target image masks.
#'
#' @param img A `magick-image`, `array` or `torch_tensor`.
#'
#' @family transforms
#'
#' @export
transform_to_tensor <- function(img) {
  UseMethod("transform_to_tensor", img)
}

#' Convert a tensor image to the given `dtype` and scale the values accordingly
#'
#' @inheritParams transform_to_tensor
#' @param dtype (torch.dtype): Desired data type of the output.
#'
#' @note
#' When converting from a smaller to a larger integer `dtype` the maximum
#'   values are **not** mapped exactly. If converted back and forth, this
#'   mismatch has no effect.
#'
#' @family transforms
#'
#' @export
transform_convert_image_dtype <- function(img, dtype = torch::torch_float()) {
  UseMethod("transform_convert_image_dtype", img)
}

#' Normalize a tensor image with mean and standard deviation
#'
#' Given mean: `(mean[1],...,mean[n])` and std: `(std[1],..,std[n])` for `n`
#'   channels, this transform will normalize each channel of the input
#'   `torch_tensor` i.e.,
#'   `output[channel] = (input[channel] - mean[channel]) / std[channel]`
#'
#' @note
#' This transform acts out of place, i.e., it does not mutate the input tensor.
#'
#' @inheritParams transform_to_tensor
#' @param mean (sequence): Sequence of means for each channel.
#' @param std (sequence): Sequence of standard deviations for each channel.
#' @param inplace (bool,optional): Bool to make this operation in-place.
#'
#' @family transforms
#'
#' @export
transform_normalize <- function(img, mean, std, inplace = FALSE) {
  UseMethod("transform_normalize", img)
}

#' Resize the input image to the given size
#'
#' The image can be a Magic Image or a torch Tensor, in which case it is
#'   expected to have `[..., H, W]` shape, where ... means an arbitrary number
#'   of leading dimensions
#'
#' @inheritParams transform_to_tensor
#' @param size (sequence or int): Desired output size. If size is a sequence
#'   like (h, w), output size will be matched to this. If size is an int,
#'   smaller edge of the image will be matched to this number.
#'   i.e, if height > width, then image will be rescaled to
#'   (size * height / width, size).
#' @param interpolation (int, optional) Desired interpolation. An integer
#'   `0 = nearest`, `2 = bilinear`, and `3 = bicubic` or a name from
#'   [magick::filter_types()].
#'
#' @family transforms
#'
#' @export
transform_resize <- function(img, size, interpolation = 2) {
  UseMethod("transform_resize", img)
}

#' Crops the given image at the center
#'
#' The image can be a Magick Image or a torch Tensor, in which case it is
#'   expected to have `[..., H, W]` shape, where ... means an arbitrary number
#'   of leading dimensions.
#'
#' @inheritParams transform_to_tensor
#' @param size (sequence or int): Desired output size of the crop. If size is
#'   an int instead of sequence like (h, w), a square crop (size, size) is
#'   made. If provided a tuple or list of length 1, it will be interpreted as
#'   `(size, size)`.
#'
#' @family transforms
#'
#' @export
transform_center_crop <- function(img, size) {
  UseMethod("transform_center_crop", img)
}

#' Pad the given image on all sides with the given "pad" value
#'
#' The image can be a Magick Image or a torch Tensor, in which case it is
#'   expected to have `[..., H, W]` shape, where ... means an arbitrary number
#'   of leading dimensions.
#'
#' @inheritParams transform_to_tensor
#' @param padding (int or tuple or list): Padding on each border. If a single
#'   int is provided this is used to pad all borders. If tuple of length 2 is
#'   provided this is the padding on left/right and top/bottom respectively.
#'   If a tuple of length 4 is provided this is the padding for the left, right,
#'   top and bottom borders respectively.
#' @param fill (int or str or tuple): Pixel fill value for constant fill.
#'   Default is 0. If a tuple of length 3, it is used to fill R, G, B channels
#'   respectively. This value is only used when the padding_mode is constant.
#'   Only int value is supported for Tensors.
#' @param padding_mode Type of padding. Should be: constant, edge, reflect or
#'   symmetric. Default is constant. Mode symmetric is not yet supported for
#'   Tensor inputs.
#'
#'  - constant: pads with a constant value, this value is specified with fill
#'
#'  - edge: pads with the last value on the edge of the image
#'
#'  - reflect: pads with reflection of image (without repeating the last
#'    value on the edge) padding `[1, 2, 3, 4]` with 2 elements on both sides
#'    in reflect mode will result in `[3, 2, 1, 2, 3, 4, 3, 2]`
#'
#'  - symmetric: pads with reflection of image (repeating the last value on
#'    the edge) padding `[1, 2, 3, 4]` with 2 elements on both sides in
#'    symmetric mode will result in `[2, 1, 1, 2, 3, 4, 4, 3]`
#'
#' @family transforms
#'
#' @export
transform_pad <- function(img, padding, fill = 0, padding_mode = "constant") {
  UseMethod("transform_pad", img)
}

#' Apply a list of transformations randomly with a given probability
#'
#' @inheritParams transform_to_tensor
#' @param transforms (list or tuple): list of transformations.
#' @param p (float): probability.
#'
#' @family transforms
#'
#' @export
transform_random_apply <- function(img, transforms, p = 0.5) {
  UseMethod("transform_random_apply", img)
}

#' Apply single transformation randomly picked from a list
#'
#' @inheritParams transform_random_apply
#'
#' @family transforms
#'
#' @export
transform_random_choice <- function(img, transforms) {
  UseMethod("transform_random_choice", img)
}

#' Apply a list of transformations in a random order
#'
#' @inheritParams transform_random_apply
#' @family transforms
#' @export
transform_random_order <- function(img, transforms) {
  UseMethod("transform_random_order", img)
}

#' Crop the given image at a random location
#'
#' The image can be a Magick Image or a Tensor, in which case it is expected
#'   to have `[..., H, W]` shape, where ... means an arbitrary number of leading
#'   dimensions.
#'
#' @inheritParams transform_resize
#' @inheritParams transform_pad
#' @param pad_if_needed (boolean): It will pad the image if smaller than the
#'   desired size to avoid raising an exception. Since cropping is done
#'   after padding, the padding seems to be done at a random offset.
#'
#' @family transforms
#'
#' @export
transform_random_crop <- function(img, size, padding=NULL, pad_if_needed=FALSE,
                                  fill=0, padding_mode="constant") {
  UseMethod("transform_random_crop", img)
}

#' Horizontally flip an image randomly with a given probability
#'
#' Horizontally flip an image randomly with a given probability. The image can
#'   be a Magick Image or a torch Tensor, in which case it is expected to have
#'   `[..., H, W]` shape, where ... means an arbitrary number of leading
#'   dimensions
#'
#' @inheritParams transform_to_tensor
#' @param p (float): probability of the image being flipped.
#'   Default value is 0.5
#'
#' @family transforms
#' @export
transform_random_horizontal_flip <- function(img, p = 0.5) {
  UseMethod("transform_random_horizontal_flip", img)
}

#' Vertically flip an image randomly with a given probability
#'
#' The image can be a PIL Image or a torch Tensor, in which case it is expected
#'   to have `[..., H, W]` shape, where `...` means an arbitrary number of
#'   leading dimensions
#'
#' @inheritParams transform_random_horizontal_flip
#'
#' @family transforms
#' @export
transform_random_vertical_flip <- function(img, p = 0.5) {
  UseMethod("transform_random_vertical_flip", img)
}

#' Crop image to random size and aspect ratio
#'
#' Crop the given image to a random size and aspect ratio. The image can be a
#'   Magick Image or a Tensor, in which case it is expected to have
#'   `[..., H, W]` shape, where ... means an arbitrary number of leading
#'   dimensions
#'
#' @details A crop of random size (default: of 0.08 to 1.0) of the original size
#'   and a random aspect ratio (default: of 3/4 to 4/3) of the original aspect
#'   ratio is made. This crop is finally resized to given size. This is
#'   popularly used to train the Inception networks.
#'
#' @inheritParams transform_resize
#' @inheritParams transform_pad
#' @param scale (tuple of float): range of size of the origin size cropped
#' @param ratio (tuple of float): range of aspect ratio of the origin aspect
#'   ratio cropped.
#'
#' @family transforms
#' @export
transform_random_resized_crop <- function(img, size, scale=c(0.08, 1.0),
                                          ratio=c(3. / 4., 4. / 3.),
                                          interpolation=2) {
  UseMethod("transform_random_resized_crop", img)
}

#' Crop image into four corners and a central crop
#'
#' Crop the given image into four corners and the central crop. This transform
#'   returns a tuple of images and there may be a mismatch in the number of
#'   inputs and targets your Dataset returns.
#'
#' @inheritParams transform_resize
#'
#' @family transforms
#' @export
transform_five_crop <- function(img, size) {
  UseMethod("transform_five_crop", img)
}

#' Crop an image and the flipped image each into four corners and a central crop
#'
#' Crop the given image into four corners and the central crop, plus the flipped
#'   version of these (horizontal flipping is used by default). This transform
#'   returns a tuple of images and there may be a mismatch in the number of
#'   inputs and targets your Dataset returns.
#'
#' @inheritParams transform_five_crop
#' @param vertical_flip (bool): Use vertical flipping instead of horizontal
#'
#' @family transforms
#' @export
transform_ten_crop <- function(img, size, vertical_flip = FALSE) {
  UseMethod("transform_ten_crop", img)
}

#' Transform a tensor image with a square transformation matrix and a
#'   mean_vector computed offline
#'
#' Given `transformation_matrix` and `mean_vector`, will flatten the
#'   `torch_tensor` and subtract `mean_vector` from it which is then followed by
#'   computing the dot product with the transformation matrix and then reshaping
#'   the tensor to its original shape.
#'
#' @section Applications:
#'   whitening transformation: Suppose X is a column vector zero-centered data.
#'   Then compute the data covariance matrix `[D x D]` with torch.mm(X.t(), X),
#'   perform SVD on this matrix and pass it as `transformation_matrix`.
#'
#' @inheritParams transform_to_tensor
#' @param transformation_matrix (Tensor): tensor `[D x D]`, D = C x H x W.
#' @param mean_vector (Tensor): tensor [D], D = C x H x W.
#'
#' @family transforms
#' @export
transform_linear_transformation <- function(img, transformation_matrix, mean_vector) {
  UseMethod("transform_linear_transformation", img)
}

#' Randomly change the brightness, contrast and saturation of an image
#'
#' @param brightness (float or tuple of float (min, max)): How much to jitter
#'   brightness. `brightness_factor` is chosen uniformly from
#'   `[max(0, 1 - brightness), 1 + brightness]` or the given `[min, max]`.
#'   Should be non negative numbers.
#' @param contrast (float or tuple of float (min, max)): How much to jitter
#'   contrast. `contrast_factor` is chosen uniformly from
#'   `[max(0, 1 - contrast), 1 + contrast]` or the given `[min, max]`. Should
#'   be non negative numbers.
#' @param saturation (float or tuple of float (min, max)): How much to jitter
#'   saturation. `saturation_factor` is chosen uniformly from
#'   `[max(0, 1 - saturation), 1 + saturation]` or the given `[min, max]`.
#'   Should be non negative numbers.
#' @param hue (float or tuple of float (min, max)): How much to jitter hue.
#'   `hue_factor` is chosen uniformly from `[-hue, hue]` or the given
#'   `[min, max]`. Should have 0<= hue <= 0.5 or -0.5 <= min <= max <= 0.5.
#' @inheritParams transform_to_tensor
#'
#' @family transforms
#' @export
transform_color_jitter <- function(img, brightness=0, contrast=0, saturation=0, hue=0) {
  UseMethod("transform_color_jitter", img)
}

#' Rotate the image by angle
#'
#' @param degrees (sequence or float or int): Range of degrees to select from.
#'   If degrees is a number instead of sequence like (min, max), the range of
#'   degrees will be (-degrees, +degrees).
#' @param resample (int, optional): An optional resampling filter. See interpolation
#'   modes.
#' @param expand (bool, optional): Optional expansion flag. If true, expands the
#'   output to make it large enough to hold the entire rotated image. If false
#'   or omitted, make the output image the same size as the input image. Note
#'   that the expand flag assumes rotation around the center and no translation.
#' @param center (list or tuple, optional): Optional center of rotation, (x, y).
#'   Origin is the upper left corner. Default is the center of the image.
#' @param fill (n-tuple or int or float): Pixel fill value for area outside the
#'   rotated image. If int or float, the value is used for all bands
#'   respectively. Defaults to 0 for all bands. This option is only available
#'   for Pillow>=5.2.0. This option is not supported for Tensor input. Fill
#'   value for the area outside the transform in the output image is always 0.
#' @inheritParams transform_to_tensor
#'
#' @family transforms
#' @export
transform_random_rotation <- function(img, degrees, resample = 0, expand=FALSE,
                                      center=NULL, fill=NULL) {
  UseMethod("transform_random_rotation", img)
}

#' Random affine transformation of the image keeping center invariant
#'
#' @inheritParams transform_random_rotation
#' @param translate (tuple, optional): tuple of maximum absolute fraction for
#'   horizontal and vertical translations. For example translate=(a, b), then
#'   horizontal shift is randomly sampled in the range
#'   -img_width * a < dx < img_width * a and vertical shift is randomly sampled
#'   in the range -img_height * b < dy < img_height * b. Will not translate by
#'   default.
#' @param scale (tuple, optional): scaling factor interval, e.g (a, b), then
#'   scale is randomly sampled from the range a <= scale <= b. Will keep
#'   original scale by default.
#' @param shear (sequence or float or int, optional): Range of degrees to select
#'   from. If shear is a number, a shear parallel to the x axis in the range
#'   (-shear, +shear) will be applied. Else if shear is a tuple or list of 2
#'   values a shear parallel to the x axis in the range `(shear[1], shear[2])`
#'   will be applied. Else if shear is a tuple or list of 4 values, a x-axis
#'   shear in `(shear[1], shear[2])` and y-axis shear in `(shear[3], shear[4])`
#'   will be applied. Will not apply shear by default.
#' @param fillcolor (tuple or int): Optional fill color (Tuple for RGB Image and
#'   int for grayscale) for the area outside the transform in the output image
#'   (Pillow>=5.0.0). This option is not supported for Tensor input. Fill value
#'   for the area outside the transform in the output image is always 0.
#'
#' @family transforms
#' @export
transform_random_affine <- function(img, degrees, translate=NULL, scale=NULL,
                                    shear=NULL, resample=0, fillcolor=0) {
  UseMethod("transform_random_affine", img)
}

#' Convert image to grayscale
#'
#' @inheritParams transform_to_tensor
#' @param num_output_channels (int): (1 or 3) number of channels desired for
#'   output image
#'
#' @family transforms
#' @export
transform_grayscale <- function(img, num_output_channels) {
  UseMethod("transform_grayscale", img)
}

#' Randomly convert image to grayscale with a given probability
#'
#' Convert image to grayscale with a probability of `p`.
#'
#' @inheritParams transform_to_tensor
#' @param p (float): probability that image should be converted to grayscale
#'   (default 0.1).
#'
#' @family transforms
#' @export
transform_random_grayscale <- function(img, p = 0.1) {
  UseMethod("transform_random_grayscale", img)
}


#' Random perspective transformation of an image with a given probability
#'
#' Performs a random perspective transformation of the given image with a given
#'   probability
#'
#' @param distortion_scale (float): argument to control the degree of distortion
#'   and ranges from 0 to 1. Default is 0.5.
#' @param p (float): probability of the image being transformed. Default is 0.5.
#' @inheritParams transform_resize
#' @inheritParams transform_pad
#'
#' @family transforms
#' @export
transform_random_perspective <- function(img, distortion_scale=0.5, p=0.5,
                                         interpolation=2, fill=0) {
  UseMethod("transform_random_perspective", img)
}

#' Randomly selects a rectangular region in an image and erases its pixel values
#'
#' 'Random Erasing Data Augmentation' by Zhong _et al._
#'   See <https://arxiv.org/pdf/1708.04896>
#'
#' @inheritParams transform_to_tensor
#' @param p probability that the random erasing operation will be performed.
#' @param scale range of proportion of erased area against input image.
#' @param ratio range of aspect ratio of erased area.
#' @param value erasing value. Default is 0. If a single int, it is used to
#'   erase all pixels. If a tuple of length 3, it is used to erase
#'   R, G, B channels respectively.
#'   If a str of 'random', erasing each pixel with random values.
#' @param inplace boolean to make this transform inplace. Default set to FALSE.
#'
#' @family transforms
#' @export
transform_random_erasing <- function(img, p=0.5, scale=c(0.02, 0.33), ratio=c(0.3, 3.3),
                                     value=0, inplace=FALSE) {
  UseMethod("transform_random_erasing", img)
}

# Other generics ----------------------------------------------------------

#' Crop the given image at specified location and output size
#'
#' @inheritParams transform_to_tensor
#' @param top (int): Vertical component of the top left corner of the crop box.
#' @param left (int): Horizontal component of the top left corner of the crop
#'   box.
#' @param height (int): Height of the crop box.
#' @param width (int): Width of the crop box.
#'
#' @family transforms
#' @export
transform_crop <- function(img, top, left, height, width) {
  UseMethod("transform_crop", img)
}

#' Horizontally flip a PIL Image or Tensor
#'
#' @inheritParams transform_to_tensor
#'
#' @family transforms
#' @export
transform_hflip <- function(img) {
  UseMethod("transform_hflip", img)
}

#' Vertically flip a PIL Image or Tensor
#'
#' @inheritParams transform_to_tensor
#'
#' @family transforms
#' @export
transform_vflip <- function(img) {
  UseMethod("transform_vflip", img)
}

#' Crop an image and resize it to a desired size
#'
#' @param top (int): Vertical component of the top left corner of the crop box.
#' @param left (int): Horizontal component of the top left corner of the crop
#'   box.
#' @param height (int): Height of the crop box.
#' @param width (int): Width of the crop box.
#' @inheritParams transform_resize
#'
#' @family transforms
#' @export
transform_resized_crop <- function(img, top, left, height, width, size,
                                   interpolation = 2) {
  UseMethod("transform_resized_crop", img)
}


#' Adjust the brightness of an image
#'
#' @param brightness_factor (float):  How much to adjust the brightness. Can be
#'   any non negative number. 0 gives a black image, 1 gives the
#'   original image while 2 increases the brightness by a factor of 2.
#' @inheritParams transform_resize
#'
#' @family transforms
#' @export
transform_adjust_brightness <- function(img, brightness_factor) {
  UseMethod("transform_adjust_brightness", img)
}

#' Adjust the contrast of an image
#'
#' @param contrast_factor (float): How much to adjust the contrast. Can be any
#'   non negative number. 0 gives a solid gray image, 1 gives the
#'   original image while 2 increases the contrast by a factor of 2.
#'
#' @inheritParams transform_resize
#'
#' @family transforms
#' @export
transform_adjust_contrast <- function(img, contrast_factor) {
  UseMethod("transform_adjust_contrast", img)
}

#' Adjust the color saturation of an image
#'
#' @param saturation_factor (float):  How much to adjust the saturation. 0 will
#'   give a black and white image, 1 will give the original image while
#'   2 will enhance the saturation by a factor of 2.
#'
#' @inheritParams transform_resize
#'
#' @family transforms
#' @export
transform_adjust_saturation <- function(img, saturation_factor) {
  UseMethod("transform_adjust_saturation", img)
}

#' Adjust the hue of an image
#'
#' The image hue is adjusted by converting the image to HSV and
#'   cyclically shifting the intensities in the hue channel (H).
#'   The image is then converted back to original image mode.
#'
#' `hue_factor` is the amount of shift in H channel and must be in the
#'   interval `[-0.5, 0.5]`.
#'
#' See [Hue](https://en.wikipedia.org/wiki/Hue) for more details.
#'
#' @param hue_factor (float):  How much to shift the hue channel. Should be in
#'   `[-0.5, 0.5]`. 0.5 and -0.5 give complete reversal of hue channel in
#'   HSV space in positive and negative direction respectively.
#'   0 means no shift. Therefore, both -0.5 and 0.5 will give an image
#'   with complementary colors while 0 gives the original image.
#'
#' @inheritParams transform_resize
#'
#' @family transforms
#' @export
transform_adjust_hue <- function(img, hue_factor) {
  UseMethod("transform_adjust_hue", img)
}

#' Angular rotation of an image
#'
#' @inheritParams transform_to_tensor
#' @inheritParams transform_random_rotation
#' @param angle (float or int): rotation angle value in degrees,
#'   counter-clockwise.
#'
#' @family transforms
#' @export
transform_rotate <- function(img, angle, resample = 0, expand = FALSE,
                             center = NULL, fill = NULL) {
  UseMethod("transform_rotate", img)
}

#' Apply affine transformation on an image keeping image center invariant
#'
#' @inheritParams transform_random_affine
#' @inheritParams transform_rotate
#'
#' @family transforms
#' @export
transform_affine <- function(img, angle, translate, scale, shear,
                             resample = 0, fillcolor = NULL) {
  UseMethod("transform_affine", img)
}

#' Perspective transformation of an image
#'
#' @param startpoints (list of list of ints): List containing four lists of two
#'   integers corresponding to four corners
#'   `[top-left, top-right, bottom-right, bottom-left]` of the original image.
#' @param endpoints (list of list of ints): List containing four lists of two
#'   integers corresponding to four corners
#'   `[top-left, top-right, bottom-right, bottom-left]` of the transformed
#'   image.
#' @inheritParams transform_resize
#' @inheritParams transform_pad
#'
#' @family transforms
#' @export
transform_perspective <- function(img, startpoints, endpoints, interpolation = 2,
                                  fill = NULL) {
  UseMethod("transform_perspective", img)
}

#' Adjust the gamma of an RGB image
#'
#' Also known as Power Law Transform. Intensities in RGB mode are adjusted
#'   based on the following equation:
#'   \deqn{
#'     I_{\mbox{out}} = 255 \times \mbox{gain} \times \left
#'     (\frac{I_{\mbox{in}}}{255}\right)^{\gamma}
#'     }
#'
#' @details See [Gamma Correction](https://en.wikipedia.org/wiki/Gamma_correction) for more details.
#'
#' @param gamma (float): Non negative real number, same as \eqn{\gamma} in the
#'   equation. gamma larger than 1 make the shadows darker, while gamma smaller
#'   than 1 make dark regions lighter.
#' @param gain (float): The constant multiplier.
#' @inheritParams transform_to_tensor
#'
#' @family transforms
#' @export
transform_adjust_gamma <- function(img, gamma, gain = 1) {
  UseMethod("transform_adjust_gamma", img)
}

#' Convert RGB Image Tensor to Grayscale
#'
#' For RGB to Grayscale conversion, ITU-R 601-2 luma transform is performed
#'   which is L = R * 0.2989 + G * 0.5870 + B * 0.1140
#'
#' @inheritParams transform_to_tensor
#'
#' @family transforms
#' @export
transform_rgb_to_grayscale <- function(img) {
  UseMethod("transform_rgb_to_grayscale")
}


# Utils -------------------------------------------------------------------

get_image_size <- function(img) {
  UseMethod("get_image_size")
}