Reliability-based Mesh-to-Grid Image Reconstruction

TL;DR

This paper introduces a reliability-based, content-adaptive image reconstruction method from mesh samples, significantly improving quality over initial estimates and outperforming existing denoising refinement techniques.

Contribution

A new reliability-based framework for mesh-to-grid image reconstruction that adaptively refines estimates using denoising, enhancing accuracy in applications like super-resolution and virtual view generation.

Findings

Improves reconstruction quality by over 2 dB PSNR.

Outperforms state-of-the-art denoising refinement by up to 0.7 dB.

Effective for applications involving non-integer sample positions.

Abstract

This paper presents a novel method for the reconstruction of images from samples located at non-integer positions, called mesh. This is a common scenario for many image processing applications, such as super-resolution, warping or virtual view generation in multi-camera systems. The proposed method relies on a set of initial estimates that are later refined by a new reliability-based content-adaptive framework that employs denoising in order to reduce the reconstruction error. The reliability of the initial estimate is computed so stronger denoising is applied to less reliable estimates. The proposed technique can improve the reconstruction quality by more than 2 dB (in terms of PSNR) with respect to the initial estimate and it outperforms the state-of-the-art denoising-based refinement by up to 0.7 dB.