Personal tools
You are here: Home Publications Regularizing Image Reconstruction for Gradient-Domain Rendering with Feature Patches

Regularizing Image Reconstruction for Gradient-Domain Rendering with Feature Patches

Computer Graphics Forum (Proceedings of Eurographics 2016)

Feature patches for gradient reconstruction

 

Marco Manzi Delio Vicini Matthias Zwicker

University of Bern

Abstract

We present a novel algorithm to reconstruct high-quality images from sampled pixels and gradients in gradient-domain rendering. Our approach extends screened Poisson reconstruction by adding additional regularization constraints. Our key idea is to exploit local patches in feature images, which contain per-pixels normals, textures, position, etc., to formulate these constraints. We describe a GPU implementation of our approach that runs on the order of seconds on megapixel images. We demonstrate a significant improvement in image quality over screened Poisson reconstruction under the L1 norm. Because we adapt the regularization constraints to the noise level in the input, our algorithm is consistent and converges to the ground truth.

Materials

publication, results and comparisons, presentation (coming soon)

Document Actions