This task type uses CLIP to mask a grayscale input image using the binary segments of the second input image.
