Generative Texture Filtering

arXiv cs.CV / 4/22/2026

📰 NewsDeveloper Stack & InfrastructureModels & Research

共有:

Key Points

The paper introduces a generative approach to texture filtering that achieves strong performance and good generalizability.
It improves texture filtering by leveraging the learned image priors from pre-trained generative models and fine-tuning them in a two-stage process.
Stage one uses supervised fine-tuning on a small set of paired images, while stage two uses reinforcement fine-tuning on a large unlabeled dataset guided by a reward function measuring texture removal quality and structure preservation.
Experiments indicate the method outperforms prior techniques and works well even on previously difficult texture-filtering cases.
The authors provide released code via GitHub, supporting reproducibility and adoption.

Abstract

We present a generative method for texture filtering, which exhibits surprisingly good performance and generalizability. Our core idea is to empower texture filtering by taking full advantage of the strong learned image prior of pre-trained generative models. To this end, we propose to fine-tune a pre-trained generative model via a two-stage strategy. Specifically, we first conduct supervised fine-tuning on a very small set of paired images, and then perform reinforcement fine-tuning on a large-scale unlabeled dataset under the guidance of a reward function that quantifies the quality of texture removal and structure preservation. Extensive experiments show that our method clearly outperforms previous methods, and is effective to deal with previously challenging cases. Our code is available at https://github.com/OnlyZZZZ/Generative_Texture_Filtering.