IEEE Transactions on Computational Imaging, 2017

A Bayesian Hyperprior Approach for Joint Image Denoising and Interpolation, with an Application to HDR Imaging
Cecilia Aguerrebere1, Andrés Almansa2, Julie Delon3, Yann Gousseau4 and Pablo Musé5

1Duke University
2LTCI - Laboratoire Traitement et Communication de l'Information
3Laboratoire MAP5 (UMR CNRS 8145), Université Paris Descartes, Sorbonne Paris Cité
4Télécom ParisTech
5 IIE - Instituto de Ingeniería Eléctrica

Stochastic Film Grain Generation
Real data. Left: Tone mapped version of the HDR image obtained by the proposed approach and its corresponding mask of unknown (black) and well-exposed (white) pixels. Right: Comparison of the results obtained by the proposed approach (first row) and PLEV (second row) in the extracts indicated in the top image. Please see the digital copy for better details reproduction.


Recently, impressive denoising results have been achieved by Bayesian approaches which assume Gaussian models for the image patches. This improvement in performance can be attributed to the use of per-patch models. Unfortunately such an approach is particularly unstable for most inverse problems beyond denoising. In this work, we propose the use of a hyperprior to model image patches, in order to stabilize the estimation procedure. There are two main advantages to the proposed restoration scheme: Firstly it is adapted to diagonal degradation matrices, and in particular to missing data problems (e.g. inpainting of missing pixels or zooming). Secondly it can deal with signal dependent noise models, particularly suited to digital cameras. As such, the scheme is especially adapted to computational photography. In order to illustrate this point, we provide an application to high dynamic range imaging from a single image taken with a modi?ed sensor, which shows the effectiveness of the proposed scheme.