Probabilistic Semantic Inpainting with Pixel Constrained CNNs
This addresses the need for diverse image inpainting in computer vision, though it is incremental as it builds on existing PixelCNN methods.
The paper tackles the problem of generating multiple plausible inpainted images for missing regions in images by proposing a probabilistic method based on PixelCNNs, which learns a distribution conditioned on visible pixels, and demonstrates diverse and realistic results on MNIST and CelebA datasets.
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.