Learning Reaction-Diffusion Models for Image Inpainting

Wei Yu, Stefan Heber, Thomas Pock

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper we present a trained diffusion model for image inpainting based on the structural similarity measure. The proposed diffusion model uses several parametrized linear filters and influence functions. Those parameters are learned in a loss based approach, where we first perform a greedy training before conducting a joint training to further improve the inpainting performance. We provide a detailed comparison to state-of-the-art inpainting algorithms based on the TUM-image inpainting database. The experimental results show that the proposed diffusion model is efficient and achieves superior performance. Moreover, we also demonstrate that the proposed method has a texture preserving property, that makes it stand out from previous PDE based methods.
LanguageEnglish
Title of host publicationPattern Recognition
Subtitle of host publication37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings
PublisherSpringer International Publishing AG
Pages356-367
Volume9358
ISBN (Electronic)978-3-319-24947-6
ISBN (Print)978-3-319-24946-9
DOIs
StatusAccepted/In press - 2015

Fingerprint

Textures

Fields of Expertise

  • Information, Communication & Computing

Cite this

Yu, W., Heber, S., & Pock, T. (2015). Learning Reaction-Diffusion Models for Image Inpainting. In Pattern Recognition: 37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings (Vol. 9358, pp. 356-367). Springer International Publishing AG . DOI: 10.1007/978-3-319-24947-6_29

Learning Reaction-Diffusion Models for Image Inpainting. / Yu, Wei; Heber, Stefan; Pock, Thomas.

Pattern Recognition: 37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings. Vol. 9358 Springer International Publishing AG , 2015. p. 356-367.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Yu, W, Heber, S & Pock, T 2015, Learning Reaction-Diffusion Models for Image Inpainting. in Pattern Recognition: 37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings. vol. 9358, Springer International Publishing AG , pp. 356-367. DOI: 10.1007/978-3-319-24947-6_29
Yu W, Heber S, Pock T. Learning Reaction-Diffusion Models for Image Inpainting. In Pattern Recognition: 37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings. Vol. 9358. Springer International Publishing AG . 2015. p. 356-367. Available from, DOI: 10.1007/978-3-319-24947-6_29
Yu, Wei ; Heber, Stefan ; Pock, Thomas. / Learning Reaction-Diffusion Models for Image Inpainting. Pattern Recognition: 37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings. Vol. 9358 Springer International Publishing AG , 2015. pp. 356-367
@inproceedings{82c7e803c7354e299b958c978ae066d9,
title = "Learning Reaction-Diffusion Models for Image Inpainting",
abstract = "In this paper we present a trained diffusion model for image inpainting based on the structural similarity measure. The proposed diffusion model uses several parametrized linear filters and influence functions. Those parameters are learned in a loss based approach, where we first perform a greedy training before conducting a joint training to further improve the inpainting performance. We provide a detailed comparison to state-of-the-art inpainting algorithms based on the TUM-image inpainting database. The experimental results show that the proposed diffusion model is efficient and achieves superior performance. Moreover, we also demonstrate that the proposed method has a texture preserving property, that makes it stand out from previous PDE based methods.",
author = "Wei Yu and Stefan Heber and Thomas Pock",
year = "2015",
doi = "10.1007/978-3-319-24947-6_29",
language = "English",
isbn = "978-3-319-24946-9",
volume = "9358",
pages = "356--367",
booktitle = "Pattern Recognition",
publisher = "Springer International Publishing AG",
address = "Switzerland",

}

TY - GEN

T1 - Learning Reaction-Diffusion Models for Image Inpainting

AU - Yu,Wei

AU - Heber,Stefan

AU - Pock,Thomas

PY - 2015

Y1 - 2015

N2 - In this paper we present a trained diffusion model for image inpainting based on the structural similarity measure. The proposed diffusion model uses several parametrized linear filters and influence functions. Those parameters are learned in a loss based approach, where we first perform a greedy training before conducting a joint training to further improve the inpainting performance. We provide a detailed comparison to state-of-the-art inpainting algorithms based on the TUM-image inpainting database. The experimental results show that the proposed diffusion model is efficient and achieves superior performance. Moreover, we also demonstrate that the proposed method has a texture preserving property, that makes it stand out from previous PDE based methods.

AB - In this paper we present a trained diffusion model for image inpainting based on the structural similarity measure. The proposed diffusion model uses several parametrized linear filters and influence functions. Those parameters are learned in a loss based approach, where we first perform a greedy training before conducting a joint training to further improve the inpainting performance. We provide a detailed comparison to state-of-the-art inpainting algorithms based on the TUM-image inpainting database. The experimental results show that the proposed diffusion model is efficient and achieves superior performance. Moreover, we also demonstrate that the proposed method has a texture preserving property, that makes it stand out from previous PDE based methods.

U2 - 10.1007/978-3-319-24947-6_29

DO - 10.1007/978-3-319-24947-6_29

M3 - Conference contribution

SN - 978-3-319-24946-9

VL - 9358

SP - 356

EP - 367

BT - Pattern Recognition

PB - Springer International Publishing AG

ER -