imcropbox

imcropbox(img: np.ndarray, box: Union[Box, np.ndarray], use_pad: bool = False) -> np.ndarray

• Description: Crop the input image using the provided bounding box.

• Parameters:

  • img (np.ndarray): Input image to be cropped.
  • box (Union[Box, np.ndarray]): Cropping box. Input can be a Box object customized by DocsaidKit, defined by coordinates (x1, y1, x2, y2), or a NumPy array with the same format.
  • use_pad (bool): Whether to use padding to handle areas outside the boundaries. If set to True, the outer regions will be padded with zeros. Default is False.

• Returns:

  • np.ndarray: Cropped image.

• Example:

  import docsaidkit as D
  
  # 使用自定義 Box 物件
  img = D.imread('lena.png')
  box = D.Box([50, 50, 200, 200], box_mode='xyxy')
  cropped_img = D.imcropbox(img, box, use_pad=True)
  
  # Resize the cropped image to the original size for visualization
  cropped_img = D.imresize(cropped_img, [img.shape[0], img.shape[1]])