def random_short_side_scale_jitter()

in datasets/video_transforms.py [0:0]

• 27 lines of code
• 8 McCabe index (conditional complexity)

def random_short_side_scale_jitter(images, min_size, max_size, boxes=None):
    """
    Perform a spatial short scale jittering on the given images and
    corresponding boxes.
    Args:
        images (tensor): images to perform scale jitter. Dimension is
            `num frames` x `channel` x `height` x `width`.
        min_size (int): the minimal size to scale the frames.
        max_size (int): the maximal size to scale the frames.
        boxes (ndarray): optional. Corresponding boxes to images.
            Dimension is `num boxes` x 4.
    Returns:
        (tensor): the scaled images with dimension of
            `num frames` x `channel` x `new height` x `new width`.
        (ndarray or None): the scaled boxes with dimension of
            `num boxes` x 4.
    """
    size = int(round(np.random.uniform(min_size, max_size)))

    height = images.shape[2]
    width = images.shape[3]
    if (width <= height and width == size) or (
        height <= width and height == size
    ):
        return images, boxes
    new_width = size
    new_height = size
    if width < height:
        new_height = int(math.floor((float(height) / width) * size))
        if boxes is not None:
            boxes = boxes * float(new_height) / height
    else:
        new_width = int(math.floor((float(width) / height) * size))
        if boxes is not None:
            boxes = boxes * float(new_width) / width

    return (
        torch.nn.functional.interpolate(
            images,
            size=(new_height, new_width),
            mode="bilinear",
            align_corners=False,
        ),
        boxes,
    )