Self2Self+: Single-Image Denoising with Self-Supervised Learning and Image Quality Assessment Loss

TL;DR

This paper introduces Self2Self+ a self-supervised single-image denoising method that leverages image quality assessment and dropout sampling to achieve state-of-the-art results without external datasets.

Contribution

The paper presents a novel self-supervised denoising approach using only a single noisy image, guided by no-reference image quality assessment and dropout sampling.

Findings

Achieved state-of-the-art denoising performance on synthetic datasets.

Effective on real-world noisy images.

No need for external noisy-clean image pairs.

Abstract

Recently, denoising methods based on supervised learning have exhibited promising performance. However, their reliance on external datasets containing noisy-clean image pairs restricts their applicability. To address this limitation, researchers have focused on training denoising networks using solely a set of noisy inputs. To improve the feasibility of denoising procedures, in this study, we proposed a single-image self-supervised learning method in which only the noisy input image is used for network training. Gated convolution was used for feature extraction and no-reference image quality assessment was used for guiding the training process. Moreover, the proposed method sampled instances from the input image dataset using Bernoulli sampling with a certain dropout rate for training. The corresponding result was produced by averaging the generated predictions from various instances of the trained network with dropouts. The experimental results indicated that the proposed method achieved state-of-the-art denoising performance on both synthetic and real-world datasets. This highlights the effectiveness and practicality of our method as a potential solution for various noise removal tasks.