Adaptive Dynamic Dehazing via Instruction-Driven and Task-Feedback Closed-Loop Optimization for Diverse Downstream Task Adaptation

2026-02-28

AI Summary

Existing image dehazing methods struggle to simultaneously enhance visual quality and meet the diverse requirements of downstream vision tasks, largely due to their inability to adapt dynamically during inference. This work proposes an adaptive dynamic dehazing framework that integrates natural language instructions and real-time feedback from downstream task performance through a closed-loop optimization mechanism, enabling on-the-fly adjustment of dehazing outputs without retraining the model. By introducing, for the first time, a dual-guidance strategy combining task-driven feedback and textual prompts, the method establishes an interactive, post-training paradigm for adaptive dehazing behavior aligned with specific task objectives. Experiments demonstrate that the proposed framework consistently improves downstream task performance across multiple vision applications while preserving high-quality dehazing results, exhibiting strong generalization and robustness.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Learning to Search
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

In real-world vision systems,haze removal is required not only to enhance image visibility but also to meet the specific needs of diverse downstream tasks.To address this challenge,we propose a novel adaptive dynamic dehazing framework that incorporates a closed-loop optimization mechanism.It enables feedback-driven refinement based on downstream task performance and user instruction-guided adjustment during inference,allowing the model to satisfy the specific requirements of multiple downstream tasks without retraining.Technically,our framework integrates two complementary and innovative mechanisms: (1)a task feedback loop that dynamically modulates dehazing outputs based on performance across multiple downstream tasks,and (2) a text instruction interface that allows users to specify high-level task preferences.This dual-guidance strategy enables the model to adapt its dehazing behavior after training,tailoring outputs in real time to the evolving needs of multiple tasks.Extensive experiments across various vision tasks demonstrate the strong effectiveness,robustness,and generalizability of our approach.These results establish a new paradigm for interactive,task-adaptive dehazing that actively collaborates with downstream applications.

Problem

• dehazing
• downstream task adaptation
• adaptive optimization
• closed-loop feedback
• instruction-driven

Innovation

• adaptive dehazing
• closed-loop optimization
• task feedback
• instruction-driven
• downstream task adaptation