dgs.utils.image.create_mask_from_polygons

dgs.utils.image.create_mask_from_polygons(img_size: tuple[int, int], polygons_x: list[list[int | float]], polygons_y: list[list[int | float]], device: torch.types.Device | str = 'cpu') → torchvision.tv_tensors.Mask[source]

Given the x- and y- coordinates of one or multiple polygons, create a Mask.

Parameters:

  • img_size – The size of the (original) image as (height, width). Because the polygon coordinates are given with respect to the original image, the returned mask will have the same shape as the image.

  • polygons_x – A list containing the x-coordinates of N polygons. Each of the polygons can have a different number of coordinates / vertices.

  • polygons_y – A list containing the y-coordinates of N polygons. Each of the polygons can have a different number of coordinates / vertices.

  • device – The device the resulting mask should be on. Default “cpu”.

Returns:

mask_tv is a binary mask containing True, where the polygons are filled. The Mask has the shape [W x H].