A method for labeling connected tubular objects within segmented image data, including: receiving segmented image data; and labeling the segmented image data to identify a plurality of components in the segmented image data, wherein the labeling includes: processing the segmented image data to create a processed image that represents centerline and radii estimates of the connected tubular components; determining seed point candidates in the processed image that are within a band of radii; grouping the candidates based on their physical distance from each other and their radii estimates; partitioning the segmented image data in accordance with the grouped candidates; and assigning a separate color label to each of the plurality of components that are different from each other.