Probabilistic Semantic Inpainting with Pixel Constrained CNNs

Emilien Dupont, Suhas Suresha
Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics, PMLR 89:2261-2270, 2019.

Abstract

Semantic inpainting is the task of inferring missing pixels in an image given surrounding pixels and high level image semantics. Most semantic inpainting algorithms are deterministic: given an image with missing regions, a single inpainted image is generated. However, there are often several plausible inpaintings for a given missing region. In this paper, we propose a method to perform probabilistic semantic inpainting by building a model, based on PixelCNNs, that learns a distribution of images conditioned on a subset of visible pixels. Experiments on the MNIST and CelebA datasets show that our method produces diverse and realistic inpaintings.

Cite this Paper


BibTeX
@InProceedings{pmlr-v89-dupont19a, title = {Probabilistic Semantic Inpainting with Pixel Constrained CNNs}, author = {Dupont, Emilien and Suresha, Suhas}, booktitle = {Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics}, pages = {2261--2270}, year = {2019}, editor = {Chaudhuri, Kamalika and Sugiyama, Masashi}, volume = {89}, series = {Proceedings of Machine Learning Research}, month = {16--18 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v89/dupont19a/dupont19a.pdf}, url = {https://proceedings.mlr.press/v89/dupont19a.html}, abstract = {Semantic inpainting is the task of inferring missing pixels in an image given surrounding pixels and high level image semantics. Most semantic inpainting algorithms are deterministic: given an image with missing regions, a single inpainted image is generated. However, there are often several plausible inpaintings for a given missing region. In this paper, we propose a method to perform probabilistic semantic inpainting by building a model, based on PixelCNNs, that learns a distribution of images conditioned on a subset of visible pixels. Experiments on the MNIST and CelebA datasets show that our method produces diverse and realistic inpaintings.} }
Endnote
%0 Conference Paper %T Probabilistic Semantic Inpainting with Pixel Constrained CNNs %A Emilien Dupont %A Suhas Suresha %B Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Masashi Sugiyama %F pmlr-v89-dupont19a %I PMLR %P 2261--2270 %U https://proceedings.mlr.press/v89/dupont19a.html %V 89 %X Semantic inpainting is the task of inferring missing pixels in an image given surrounding pixels and high level image semantics. Most semantic inpainting algorithms are deterministic: given an image with missing regions, a single inpainted image is generated. However, there are often several plausible inpaintings for a given missing region. In this paper, we propose a method to perform probabilistic semantic inpainting by building a model, based on PixelCNNs, that learns a distribution of images conditioned on a subset of visible pixels. Experiments on the MNIST and CelebA datasets show that our method produces diverse and realistic inpaintings.
APA
Dupont, E. & Suresha, S.. (2019). Probabilistic Semantic Inpainting with Pixel Constrained CNNs. Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 89:2261-2270 Available from https://proceedings.mlr.press/v89/dupont19a.html.

Related Material