eduzhai > Applied Sciences > Engineering >

Nonlocal Adaptive Direction-Guided Structure Tensor Total Variation For Image Recovery

  • king
  • (0) Download
  • 20210507
  • Save

... pages left unread,continue reading

Document pages: 9 pages

Abstract: A common strategy in variational image recovery is utilizing the nonlocalself-similarity (NSS) property, when designing energy functionals. One suchcontribution is nonlocal structure tensor total variation (NLSTV), which liesat the core of this study. This paper is concerned with boosting the NLSTVregularization term through the use of directional priors. More specifically,NLSTV is leveraged so that, at each image point, it gains more sensitivity inthe direction that is presumed to have the minimum local variation. The actualdifficulty here is capturing this directional information from the corruptedimage. In this regard, we propose a method that employs anisotropic Gaussiankernels to estimate directional features to be later used by our proposedmodel. The experiments validate that our entire two-stage framework achievesbetter results than the NLSTV model and two other competing local models, interms of visual and quantitative evaluation.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...