pad

pad(img: np.ndarray, pad_size: int | tuple[int, int] | tuple[int, int, int, int], pad_value: int | tuple[int, ...] | None = 0, pad_mode: str | int | BORDER = BORDER.CONSTANT) -> np.ndarray

Description: Applies padding to the input image.
Parameters
- img (np.ndarray): The input image to be padded.
- pad_size (Union[int, Tuple[int, int], Tuple[int, int, int, int]]): The padding size. Can be an integer to specify the same padding for all sides, a tuple (pad_top, pad_bottom, pad_left, pad_right) for different padding sizes for each side, or a tuple (pad_height, pad_width) for the same padding size for height and width.
- pad_value (int | tuple[int, ...] | None): Padding value. For 3-channel images, you can pass an int or a tuple (OpenCV convention: BGR). For grayscale images, it must be an int. Default is 0.
- pad_mode (Union[str, int, BORDER]): The padding mode. Available options:
  - BORDER.CONSTANT: Pad with a constant value (pad_value).
  - BORDER.REPLICATE: Pad by replicating the edge pixels.
  - BORDER.REFLECT: Pad by reflecting the image around the edge.
  - BORDER.REFLECT_101: Pad by reflecting the image around the edge with a slight adjustment to avoid artificial artifacts. Default is BORDER.CONSTANT.
Returns
- np.ndarray: The padded image.

Example

import capybara as cb

img = cb.imread('lena.png')
pad_img = cb.pad(img, pad_size=20, pad_value=(255, 0, 0))

# Resize the padded image to the original size for visualization
pad_img = cb.imresize(pad_img, [img.shape[0], img.shape[1]])

pad