Quantitative Assessment of Image Segmentation Quality by Random Walk Relaxation Times
The purpose of image segmentation is to partition the pixel grid of an image into connected components termed segments such that (i) each segment is homogenous and (ii) for any pair of adjacent segments, their union is not homogenous. (If it were homogenous the segments should be merged). We propose a rigorous definition of segment homogeneity which is scale-free and adaptive to the geometry of segments. We motivate this definition using random walk theory and show how segment homogeneity facilitates the quantification of violations of the conditions (i) and (ii) which are referred to as under-segmentation and over-segmentation, respectively. We describe the theoretical foundations of our approach and present a proof of concept on a few natural images.
Unable to display preview. Download preview PDF.
- 5.Grady, L., Sinop, A.K.: A seeded image segmentation framework unifying graph cuts and random walker which yields a new algorithm. In: ICCV. IEEE, Los Alamitos (2007)Google Scholar
- 6.Grady, L., Sinop, A.K.: Fast approximate random walker segmentation using eigenvector precomputation. In: CVPR, IEEE, Los Alamitos (2008)Google Scholar
- 8.Martin, D., Fowlkes, C., Tal, D., Malik, J.: A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics. In: ICCV, vol. 2, pp. 416–423 (2001)Google Scholar
- 9.Meila, M., Shi, J.: Learning segmentation by random walks. In: NIPS, pp. 873–879. MIT Press, Cambridge (2000)Google Scholar
- 10.Nadler, B., Lafon, S., Kevrekidis, I., Coifman, R.: Diffusion maps, spectral clustering and eigenfunctions of fokker-planck operators. In: NIPS, pp. 955–962. MIT Press, Cambridge (2005)Google Scholar