Randomly apply an affine transformation on a dataset item

Draws a random rotation, translation, scale and shear inside the given ranges and applies the resulting affine transformation to the dataset item with item_transform_affine, keeping the image size unchanged. Image and target share the same draw, so that they stay aligned.

item_transform_random_affine(
  x,
  degrees,
  translate = NULL,
  scale = NULL,
  shear = NULL,
  interpolation = 0,
  fill = NULL,
  center = NULL
)

Arguments

x

A dataset item, typically an image_with_bounding_box, image_with_rotated_box or image_with_segmentation_mask object containing an image tensor and associated target data.

degrees

(numeric vector of length 2 or numeric): Range of degrees to select from. If degrees is a bare number instead of a numeric vector of length 2 like c(min, max), the range of degrees will be (-degrees, +degrees).

translate

(numeric vector of length 2, optional): Numeric vector of maximum absolute fraction for horizontal and vertical translations. For example translate = c(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.

scale

(numeric vector of length 2, optional): Scaling factor interval, e.g. c(a, b), then scale is randomly sampled from the range a <= scale <= b. Will keep original scale by default.

shear

(numeric vector or numeric, optional): Range of degrees to select from. If shear is a bare number, a shear parallel to the x axis in the range (-shear, +shear) will be applied. Else if shear is a numeric vector of length 2, a shear parallel to the x axis in the range (shear[1], shear[2]) will be applied. Else if shear is a numeric vector of length 4, 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.

interpolation

(integer or character, optional): Interpolation mode. Supported values are 0 / "nearest" and 2 / "bilinear". Default is 0.

fill

Fill color for the area outside the transform. Default is NULL.

center

(numeric vector of length 2, optional): Optional center of rotation, c(x, y). Default is image center.

Value

A dataset item with the image and target transformed. Detection items are returned as image_with_rotated_box; segmentation items keep their class.

Details

The transformation is drawn again for every item, so that a dataset wrapped with this transform yields a different transformation on each access.

See also

Other item_random_transforms: item_transform_random_crop(), item_transform_random_erasing(), item_transform_random_horizontal_flip(), item_transform_random_perspective(), item_transform_random_resize_crop(), item_transform_random_rotation(), item_transform_random_vertical_flip()

Examples

if (FALSE) { # \dontrun{
url <- "https://upload.wikimedia.org/wikipedia/commons/b/b6/Felis_catus-cat_on_snow.jpg"
img <- base_loader(url) |> transform_to_tensor()

boxes <- torch_tensor(matrix(c(600, 200, 2880, 1860), ncol = 4), dtype = torch_float32())

before <- list(x = img, y = list(boxes = boxes, labels = "cat",
                                 image_height = img$shape[2], image_width = img$shape[3]))
class(before) <- c("image_with_bounding_box", "list")

after <- item_transform_random_affine(before, degrees = 30, translate = c(0.1, 0.1),
                                      scale = c(0.8, 1.2), shear = 10)

before_plot <- draw_bounding_boxes(before, colors = "blue", width = 10)
after_plot <- draw_bounding_boxes(after, colors = "red", width = 10)
tensor_image_browse(before_plot)
tensor_image_browse(after_plot)
} # }