HAL will be down for maintenance from Friday, June 10 at 4pm through Monday, June 13 at 9am. More information
Skip to Main content Skip to Navigation
Journal articles

Visualizing the Uncertainty of Graph-based 2D Segmentation with Min-path Stability

Abstract : This paper presents a novel approach to visualize the uncertainty in graph-based segmentations of scalar data. Segmentation of 2D scalar data has wide application in a variety of scientific and medical domains. Typically, a segmentation is presented as a single unambiguous boundary although the solution is often uncertain due to noise or blur in the underlying data as well as imprecision in user input. Our approach provides insight into this uncertainty by computing the " min-path stability " , a scalar measure analyzing the stability of the segmentation given a set of input constraints. Our approach is efficient, easy to compute, and can be generally applied to either graph cuts or live-wire (even partial) segmentations. In addition to its general applicability, our new approach to graph cuts uncertainty visualization improves on the time complexity of the current state-of-the-art with an additional fast approximate solution. We also introduce a novel query enabled by our approach which provides users with alternate segmentations by efficiently extracting local minima of the segmentation optimization. Finally, we evaluate our approach and demonstrate its utility on data from scientific and medical applications.
Complete list of metadata

Cited literature [33 references]  Display  Hide  Download

Contributor : Julien Tierny Connect in order to contact the contributor
Submitted on : Thursday, April 27, 2017 - 8:17:33 AM
Last modification on : Friday, January 8, 2021 - 5:40:03 PM
Long-term archiving on: : Friday, July 28, 2017 - 12:29:22 PM


Files produced by the author(s)




  • HAL Id : hal-01510992, version 1


Brian Summa, Julien Tierny, Valerio Pascucci. Visualizing the Uncertainty of Graph-based 2D Segmentation with Min-path Stability . Computer Graphics Forum, Wiley, 2017. ⟨hal-01510992⟩



Record views


Files downloads