Linkage between piecewise constant mumford-shah model and rudin-osher-fatemi model and its virtue in image segmentation

Xiaohao CAI, Raymond CHAN, Carola Bibiane SCHÖNLIEB, Gabriele STEIDL, Tieyong ZENG

Research output: Journal PublicationsJournal Article (refereed)peer-review

22 Citations (Scopus)

Abstract

The piecewise constant Mumford{Shah (PCMS) model and the Rudin{Osher{Fatemi (ROF) model are two important variational models in image segmentation and image restoration, respectively. In this paper, we explore a linkage between these models. We prove that for the twophase segmentation problem a partial minimizer of the PCMS model can be obtained by thresholding the minimizer of the ROF model. A similar linkage is still valid for multiphase segmentation under specific assumptions. Thus it opens a new segmentation paradigm: image segmentation can be done via image restoration plus thresholding. This new paradigm, which circumvents the innate nonconvex property of the PCMS model, therefore, improves the segmentation performance in both efficiency (much faster than state-of-the-art methods based on the PCMS model, particularly when the phase number is high)the and effectiveness (producing segmentation results with better quality) due to the flexibility of the ROF model in tackling degraded images, such as noisy images, blurry images, or images with information loss. As a by-product of the new paradigm, we derive a novel segmentation method, called thresholded-ROF (T-ROF) method, to illustrate the virtue of managing image segmentation through image restoration techniques. The convergence of the T-ROF method is proved, and elaborate experimental results and comparisons are presented.

Original languageEnglish
Pages (from-to)B1310-B1340
Number of pages31
JournalSIAM Journal on Scientific Computing
Volume41
Issue number6
DOIs
Publication statusPublished - 2019
Externally publishedYes

Bibliographical note

Publisher Copyright:
Copyright © by SIAM.

Keywords

  • Chan-Vese model
  • Image restoration
  • Image segmentation
  • Mumford{Shah model
  • Thresholding
  • Total variation ROF model

Fingerprint

Dive into the research topics of 'Linkage between piecewise constant mumford-shah model and rudin-osher-fatemi model and its virtue in image segmentation'. Together they form a unique fingerprint.

Cite this