Dual-Camera Super-Resolution with Aligned Attention Modules

Current smartphones use an asymmetric-cameras system consisting of multiple fixed-focal lenses. The most common configuration has dual cameras with wide-angle and telephoto lenses that have different fields of view.

A recent paper on arXiv.org proposes to use the telephoto image as a reference to enhance the resolution of the wide-angle image.

Image credit: Cristina Zaragoza/Unsplash

The researchers propose a reference-based super-resolution (RefSR) method with a focus on dual-camera super-resolution. A self-supervised domain adaptation scheme is proposed to bridge domain gaps between real-world images and downsampled images. The aligned attention module and adaptive fusion module are used to improve the RefSR architecture.

The experiments suggested method outperforms state-of-the-art approaches qualitatively and quantitatively.

We present a novel approach to reference-based super-resolution (RefSR) with the focus on dual-camera super-resolution (DCSR), which utilizes reference images for high-quality and high-fidelity results. Our proposed method generalizes the standard patch-based feature matching with spatial alignment operations. We further explore the dual-camera super-resolution that is one promising application of RefSR, and build a dataset that consists of 146 image pairs from the main and telephoto cameras in a smartphone. To bridge the domain gaps between real-world images and the training images, we propose a self-supervised domain adaptation strategy for real-world images. Extensive experiments on our dataset and a public benchmark demonstrate clear improvement achieved by our method over state of the art in both quantitative evaluation and visual comparisons.

Research paper: Wang, T., Xie, J., Sun, W., Yan, Q., and Chen, Q., "Dual-Camera Super-Resolution with Aligned Attention Modules", 2021. Link: https://arxiv.org/abs/2109.01349