Image despeckling with non-local total bounded variation regularization

Simple item page
dc.contributor.authorPadikkal, P.
dc.contributor.authorBanothu, B.
dc.date.accessioned2026-02-05T09:31:09Z
dc.date.issued2018
dc.description.abstractA non-local total bounded variational (TBV) regularization model is proposed for restoring images corrupted with data-correlated speckles and linear blurring artifacts. The energy functional of the model is derived using maximum a posteriori (MAP) estimate of the noise probability density function (PDF). The non-local total bounded variation prior regularizes the model while the data fidelity is derived using the MAP estimator of the noise PDF. The computational efficiency of the model is improved using a fast numerical scheme based on the Augmented Lagrange formulation. The proposed model is employed to restore ultrasound (US) and synthetic aperture radar (SAR) images, which are usually speckled and blurred. The numerical results are presented and compared. Furthermore, a detailed theoretical study of the model is performed in addition to the experimental analysis. © 2017 Elsevier Ltd
dc.identifier.citationComputers and Electrical Engineering, 2018, 70, , pp. 631-646
dc.identifier.issn457906
dc.identifier.urihttps://doi.org/10.1016/j.compeleceng.2017.09.013
dc.identifier.urihttps://idr.nitk.ac.in/handle/123456789/25073
dc.publisherElsevier Ltd
dc.subjectComputational efficiency
dc.subjectConstrained optimization
dc.subjectImage reconstruction
dc.subjectLagrange multipliers
dc.subjectSynthetic aperture radar
dc.subjectAugmented Lagrangians
dc.subjectDeblurring
dc.subjectGamma noise
dc.subjectNonlocal
dc.subjectRegularization
dc.subjectProbability density function
dc.titleImage despeckling with non-local total bounded variation regularization

Files

Collections

Journal Articles