Self-supervised multi-scale uniform motion deblurring via alternating optimization

Lening Guo

doi:10.1016/j.patcog.2025.112774

Self-supervised multi-scale uniform motion deblurring via alternating optimization

2025-11-19

0

PRE

AI

DOI：10.1016/j.patcog.2025.112774

Original

Original request for help

Save

Abstract

En 中文

• Propose a self-supervised method which is supervised by the only blurred image itself. • Integrate deep learning-based methods and mathematical optimization-based methods. • Design a multi-input multi-output network for collaborative multi-scale deblurring. • Jointly estimate the latent image and the blur kernel via alternating optimization.

Self-supervised multi-scale uniform motion deblurring via alternating optimization

Abstract

Journal

Organization