Self-supervised regression learning using domain knowledge: Applications to improving self-supervised denoising in imaging

TL;DR

This paper introduces a versatile self-supervised regression framework that leverages domain knowledge through pseudo-predictors, significantly enhancing denoising performance in imaging applications without requiring ground-truth data.

Contribution

It presents a general SSRL framework that incorporates domain knowledge via pseudo-predictors, improving self-supervised regression tasks like denoising.

Findings

SSRL outperforms existing self-supervised denoising methods.

Using domain knowledge improves regression learning.

Numerical results show significant denoising quality gains.

Abstract

Regression that predicts continuous quantity is a central part of applications using computational imaging and computer vision technologies. Yet, studying and understanding self-supervised learning for regression tasks - except for a particular regression task, image denoising - have lagged behind. This paper proposes a general self-supervised regression learning (SSRL) framework that enables learning regression neural networks with only input data (but without ground-truth target data), by using a designable pseudo-predictor that encapsulates domain knowledge of a specific application. The paper underlines the importance of using domain knowledge by showing that under different settings, the better pseudo-predictor can lead properties of SSRL closer to those of ordinary supervised learning. Numerical experiments for low-dose computational tomography denoising and camera image denoising demonstrate that proposed SSRL significantly improves the denoising quality over several existing self-supervised denoising methods.