InstaRevive: One-Step Image Enhancement via Dynamic Score Matching

2025-04-22

AI Summary

Existing diffusion-based image enhancement methods rely on multi-step iterative sampling, incurring substantial computational overhead. To address this, we propose InstaRevive—a single-step, high-fidelity image enhancement framework grounded in dynamic score matching. Its core innovation lies in the first integration of dynamic diffusion range control with cross-modal text guidance—derived from image captioning—to precisely model denoising trajectories. By jointly leveraging score-matching distillation, gradient-aligned distribution matching, and text-conditioned injection, InstaRevive reproduces the generation quality of multi-step diffusion within a single forward pass. Evaluated on low-light, haze, and underwater enhancement benchmarks, InstaRevive achieves state-of-the-art visual quality at significantly reduced inference cost, effectively balancing efficiency and fidelity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Image enhancement finds wide-ranging applications in real-world scenarios due to complex environments and the inherent limitations of imaging devices. Recent diffusion-based methods yield promising outcomes but necessitate prolonged and computationally intensive iterative sampling. In response, we propose InstaRevive, a straightforward yet powerful image enhancement framework that employs score-based diffusion distillation to harness potent generative capability and minimize the sampling steps. To fully exploit the potential of the pre-trained diffusion model, we devise a practical and effective diffusion distillation pipeline using dynamic control to address inaccuracies in updating direction during score matching. Our control strategy enables a dynamic diffusing scope, facilitating precise learning of denoising trajectories within the diffusion model and ensuring accurate distribution matching gradients during training. Additionally, to enrich guidance for the generative power, we incorporate textual prompts via image captioning as auxiliary conditions, fostering further exploration of the diffusion model. Extensive experiments substantiate the efficacy of our framework across a diverse array of challenging tasks and datasets, unveiling the compelling efficacy and efficiency of InstaRevive in delivering high-quality and visually appealing results. Code is available at https://github.com/EternalEvan/InstaRevive.

Problem

• Reduces iterative sampling steps in diffusion-based image enhancement
• Improves accuracy in score matching via dynamic control strategy
• Enhances generative guidance using textual prompts from image captions

Innovation

• Score-based diffusion distillation for fast enhancement
• Dynamic control for accurate score matching
• Textual prompts via image captioning as conditions