| Paper: | IMDSP-L5.4 | ||
| Session: | Image Segmentation | ||
| Time: | Wednesday, May 17, 15:00 - 15:20 | ||
| Presentation: | Lecture | ||
| Topic: | Image and Multidimensional Signal Processing: Image Segmentation | ||
| Title: | New Criteria for Evaluating Image Segmentation Results | ||
| Authors: | Sylvie Philipp-Foliguet, ETIS, France; Laurent Guigues, CREATIS, France | ||
| Abstract: | This paper is dedicated to evaluation criteria for image segmentation when there is no ground truth. Formulating the image segmentation as a piecewise modeling problem, new criteria are proposed. They take into account both the complexity of the segmented image, through the total boundary length and the goodness-of-fit through a distance between model and initial image. The main interest is avoid presetting the expected level of detail in the segmentation, leaving the choice to the users, accordingly to their purpose. These new evaluation criteria are compared to the most used evaluation criteria both on results of segmentation algorithms and on manual segmentation achieved by humans. | ||
©2018 Conference Management Services, Inc. -||- email: webmaster@icassp2006.org -||- Last updated Friday, August 17, 2012