Perception-Oriented Stereo Image Super-Resolution

TL;DR

This paper introduces a perception-oriented stereo image super-resolution method that emphasizes visual quality, utilizing a new stereo image quality assessment model and database to guide improvements beyond traditional metrics.

Contribution

It presents the first stereo image super-resolution approach focused on perceptual quality, supported by a novel stereo image quality assessment model and database.

Findings

Significant improvement in perceptual quality of stereo images

Enhanced reliability of stereo images for disparity estimation

Development of the first stereo image super-resolution quality assessment model

Abstract

Recent studies of deep learning based stereo image super-resolution (StereoSR) have promoted the development of StereoSR. However, existing StereoSR models mainly concentrate on improving quantitative evaluation metrics and neglect the visual quality of super-resolved stereo images. To improve the perceptual performance, this paper proposes the first perception-oriented stereo image super-resolution approach by exploiting the feedback, provided by the evaluation on the perceptual quality of StereoSR results. To provide accurate guidance for the StereoSR model, we develop the first special stereo image super-resolution quality assessment (StereoSRQA) model, and further construct a StereoSRQA database. Extensive experiments demonstrate that our StereoSR approach significantly improves the perceptual quality and enhances the reliability of stereo images for disparity estimation.