Triangulation cut for image segmentation

This paper proposes a novel method called triangulation cut (Tcut) to improve upon current color image segmentation techniques. Tcut utilizes constrained delaunay triangulation, scale based tuning and graph spectral segmentation. Tcut approach has superior computational and memory efficiency over Normalized Cuts (Ncut). In addition, it avoids over- segmentation of large homogeneous regions and it groups spatially discontiguous but similar regions in the same segment. This results in an accurate estimate of the number of perceivable regions present in the image as compared to that of Ncut. Application of Tcut on variety of images have yielded promising results.