A Simple Combination of Diffusion Models for Better Quality Trade-Offs in Image Denoising

2025-03-18

AI Summary

Existing diffusion models struggle to simultaneously achieve high visual quality and signal fidelity in Gaussian denoising. To address this, we propose Linearly Combined Diffusion Denoisers (LCDD), the first framework to instantiate a dual-path inference mechanism—comprising generative guidance and signal fidelity preservation—governed by a single tunable scalar hyperparameter that explicitly balances the two objectives without fine-tuning or retraining. Built upon pre-trained diffusion models, LCDD introduces noise-residual modeling and linear-weighted fusion to enable efficient, stable, and controllable inference. Extensive experiments across multiple benchmark datasets demonstrate that LCDD significantly outperforms conventional denoising methods and state-of-the-art diffusion-based approaches, achieving new SOTA performance with substantial PSNR and SSIM improvements.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Diffusion models have garnered considerable interest in computer vision, owing both to their capacity to synthesize photorealistic images and to their proven effectiveness in image reconstruction tasks. However, existing approaches fail to efficiently balance the high visual quality of diffusion models with the low distortion achieved by previous image reconstruction methods. Specifically, for the fundamental task of additive Gaussian noise removal, we first illustrate an intuitive method for leveraging pretrained diffusion models. Further, we introduce our proposed Linear Combination Diffusion Denoiser (LCDD), which unifies two complementary inference procedures - one that leverages the model's generative potential and another that ensures faithful signal recovery. By exploiting the inherent structure of the denoising samples, LCDD achieves state-of-the-art performance and offers controlled, well-behaved trade-offs through a simple scalar hyperparameter adjustment.

Problem

• Balancing visual quality and low distortion in image denoising
• Improving additive Gaussian noise removal using diffusion models
• Proposing LCDD for better trade-offs in image reconstruction

Innovation

• Combines diffusion models for image denoising
• Introduces Linear Combination Diffusion Denoiser (LCDD)
• Uses scalar hyperparameter for quality trade-offs