OpenRR-5k: A Large-Scale Benchmark for Reflection Removal in the Wild

TL;DR

This paper introduces OpenRR-5k, a large-scale, high-quality dataset for reflection removal in images, enabling better training and evaluation of reflection removal algorithms in diverse real-world scenarios.

Contribution

We created and released a comprehensive dataset with 5,300 pixel-aligned image pairs for reflection removal, addressing the lack of large-scale datasets in this domain.

Findings

A U-Net-based model trained on our dataset achieves competitive results.

The dataset covers diverse real-world reflection scenarios.

Evaluation metrics demonstrate the dataset's effectiveness for reflection removal tasks.

Abstract

Removing reflections is a crucial task in computer vision, with significant applications in photography and image enhancement. Nevertheless, existing methods are constrained by the absence of large-scale, high-quality, and diverse datasets. In this paper, we present a novel benchmark for Single Image Reflection Removal (SIRR). We have developed a large-scale dataset containing 5,300 high-quality, pixel-aligned image pairs, each consisting of a reflection image and its corresponding clean version. Specifically, the dataset is divided into two parts: 5,000 images are used for training, and 300 images are used for validation. Additionally, we have included 100 real-world testing images without ground truth (GT) to further evaluate the practical performance of reflection removal methods. All image pairs are precisely aligned at the pixel level to guarantee accurate supervision. The dataset encompasses a broad spectrum of real-world scenarios, featuring various lighting conditions, object types, and reflection patterns, and is segmented into training, validation, and test sets to facilitate thorough evaluation. To validate the usefulness of our dataset, we train a U-Net-based model and evaluate it using five widely-used metrics, including PSNR, SSIM, LPIPS, DISTS, and NIQE. We will release both the dataset and the code on https://github.com/caijie0620/OpenRR-5k to facilitate future research in this field.