Deep Image Prior with L0 Gradient Regularizer for Image Smoothing

2026-01-19

AI Summary

This work proposes an unsupervised image smoothing method that requires no training data yet effectively removes texture and fine details while preserving strong edges and structural integrity. The approach introduces, for the first time, an L0 gradient regularization term into the deep image prior framework, formulating a non-convex and non-smooth optimization model. This model is efficiently solved via the alternating direction method of multipliers (ADMM) coupled with a specialized L0 solver. Extensive experiments demonstrate that the proposed method significantly outperforms existing techniques in both edge-preserving smoothing and JPEG artifact removal, achieving high-quality, structure-aware image processing without supervision.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Non-convex Optimization
• Machine Learning: Learning on the Edge & Model Compression

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Image smoothing is a fundamental image processing operation that preserves the underlying structure, such as strong edges and contours, and removes minor details and textures in an image. Many image smoothing algorithms rely on computing local window statistics or solving an optimization problem. Recent state-of-the-art methods leverage deep learning, but they require a carefully curated training dataset. Because constructing a proper training dataset for image smoothing is challenging, we propose DIP-$\ell_0$, a deep image prior framework that incorporates the $\ell_0$ gradient regularizer. This framework can perform high-quality image smoothing without any training data. To properly minimize the associated loss function that has the nonconvex, nonsmooth $\ell_0$ ``norm", we develop an alternating direction method of multipliers algorithm that utilizes an off-the-shelf $\ell_0$ gradient minimization solver. Numerical experiments demonstrate that the proposed DIP-$\ell_0$ outperforms many image smoothing algorithms in edge-preserving image smoothing and JPEG artifact removal.

Problem

• image smoothing
• edge preservation
• training-free
• L0 gradient regularization
• deep image prior

Innovation

• Deep Image Prior
• L0 gradient regularization
• image smoothing
• ADMM
• unsupervised learning