Image Inpainting Using Nonlocal Texture Matching and Nonlinear Filtering

Ding Ding, Sundaresh Ram, Jeffrey J Rodriguez

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

Nonlocal texture similarity and local intensity smoothness are both essential for solving most image inpainting problems. In this paper, we propose a novel image inpainting algorithm that is capable of reproducing the underlying textural details using a nonlocal texture measure and also smoothing pixel intensity seamlessly in order to achieve natural-looking inpainted images. For matching texture, we propose a Gaussian-weighted nonlocal texture similarity measure to obtain multiple candidate patches for each target patch. To compute the pixel intensity, we apply the α-trimmed mean filter to the candidate patches to inpaint the target patch pixel by pixel. The proposed algorithm is compared with four current image inpainting algorithms under different scenarios including object removal, texture synthesis, and error concealment. Experimental results show that the proposed algorithm outperforms the existing algorithms when inpainting large missing regions in images with texture and geometric structures.

Original languageEnglish (US)
JournalIEEE Transactions on Image Processing
DOIs
StateAccepted/In press - Jan 1 2018

Fingerprint

Nonlinear filtering
Textures
Pixels

Keywords

  • α-trimmed mean filter
  • Exemplar-based image inpainting
  • Filtering
  • Image edge detection
  • Image reconstruction
  • Indexes
  • Kernel
  • nonlocal texture matching
  • object removal
  • Random access memory
  • Smoothing methods
  • texture synthesis

ASJC Scopus subject areas

  • Software
  • Computer Graphics and Computer-Aided Design

Cite this

Image Inpainting Using Nonlocal Texture Matching and Nonlinear Filtering. / Ding, Ding; Ram, Sundaresh; Rodriguez, Jeffrey J.

In: IEEE Transactions on Image Processing, 01.01.2018.

Research output: Contribution to journalArticle

@article{23a1e4cd91f9461ea7fe2a555fd9c167,
title = "Image Inpainting Using Nonlocal Texture Matching and Nonlinear Filtering",
abstract = "Nonlocal texture similarity and local intensity smoothness are both essential for solving most image inpainting problems. In this paper, we propose a novel image inpainting algorithm that is capable of reproducing the underlying textural details using a nonlocal texture measure and also smoothing pixel intensity seamlessly in order to achieve natural-looking inpainted images. For matching texture, we propose a Gaussian-weighted nonlocal texture similarity measure to obtain multiple candidate patches for each target patch. To compute the pixel intensity, we apply the α-trimmed mean filter to the candidate patches to inpaint the target patch pixel by pixel. The proposed algorithm is compared with four current image inpainting algorithms under different scenarios including object removal, texture synthesis, and error concealment. Experimental results show that the proposed algorithm outperforms the existing algorithms when inpainting large missing regions in images with texture and geometric structures.",
keywords = "α-trimmed mean filter, Exemplar-based image inpainting, Filtering, Image edge detection, Image reconstruction, Indexes, Kernel, nonlocal texture matching, object removal, Random access memory, Smoothing methods, texture synthesis",
author = "Ding Ding and Sundaresh Ram and Rodriguez, {Jeffrey J}",
year = "2018",
month = "1",
day = "1",
doi = "10.1109/TIP.2018.2880681",
language = "English (US)",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Image Inpainting Using Nonlocal Texture Matching and Nonlinear Filtering

AU - Ding, Ding

AU - Ram, Sundaresh

AU - Rodriguez, Jeffrey J

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Nonlocal texture similarity and local intensity smoothness are both essential for solving most image inpainting problems. In this paper, we propose a novel image inpainting algorithm that is capable of reproducing the underlying textural details using a nonlocal texture measure and also smoothing pixel intensity seamlessly in order to achieve natural-looking inpainted images. For matching texture, we propose a Gaussian-weighted nonlocal texture similarity measure to obtain multiple candidate patches for each target patch. To compute the pixel intensity, we apply the α-trimmed mean filter to the candidate patches to inpaint the target patch pixel by pixel. The proposed algorithm is compared with four current image inpainting algorithms under different scenarios including object removal, texture synthesis, and error concealment. Experimental results show that the proposed algorithm outperforms the existing algorithms when inpainting large missing regions in images with texture and geometric structures.

AB - Nonlocal texture similarity and local intensity smoothness are both essential for solving most image inpainting problems. In this paper, we propose a novel image inpainting algorithm that is capable of reproducing the underlying textural details using a nonlocal texture measure and also smoothing pixel intensity seamlessly in order to achieve natural-looking inpainted images. For matching texture, we propose a Gaussian-weighted nonlocal texture similarity measure to obtain multiple candidate patches for each target patch. To compute the pixel intensity, we apply the α-trimmed mean filter to the candidate patches to inpaint the target patch pixel by pixel. The proposed algorithm is compared with four current image inpainting algorithms under different scenarios including object removal, texture synthesis, and error concealment. Experimental results show that the proposed algorithm outperforms the existing algorithms when inpainting large missing regions in images with texture and geometric structures.

KW - α-trimmed mean filter

KW - Exemplar-based image inpainting

KW - Filtering

KW - Image edge detection

KW - Image reconstruction

KW - Indexes

KW - Kernel

KW - nonlocal texture matching

KW - object removal

KW - Random access memory

KW - Smoothing methods

KW - texture synthesis

UR - http://www.scopus.com/inward/record.url?scp=85056315704&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056315704&partnerID=8YFLogxK

U2 - 10.1109/TIP.2018.2880681

DO - 10.1109/TIP.2018.2880681

M3 - Article

C2 - 30418909

AN - SCOPUS:85056315704

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

ER -