This paper presents a containing ‘ground truth’ segmentations produced by humans for images of a wide variety of natural scenes-This paper presents a containing ‘ground truth’ segmentations produced by humans for images of a wide variety of