Iterative Predictor-Critic Code Decoding for Real-World Image Dehazing

2025-03-17

AI Summary

This work addresses the distortion in real-scene image dehazing caused by dynamic variations in haze density and scene depth. We propose an iterative code-encoder-decoder framework leveraging VQGAN codebook priors. Methodologically, we introduce a learnable Code-Critic module to model inter-code correlations for dynamic, high-confidence code selection; integrate a residual prediction network with a mask-score-driven code resampling mechanism to guide multi-round code prediction using prior decoding outputs, thereby overcoming the accuracy limitation of single-step decoding. To our knowledge, this is the first framework that deeply unifies code-space discriminative optimization with iterative resampling. Extensive experiments demonstrate significant improvements over state-of-the-art methods across multiple real-world dehazing benchmarks, particularly excelling in dense-haze and complex-depth scenarios—preserving structural integrity and color fidelity with remarkable consistency.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

We propose a novel Iterative Predictor-Critic Code Decoding framework for real-world image dehazing, abbreviated as IPC-Dehaze, which leverages the high-quality codebook prior encapsulated in a pre-trained VQGAN. Apart from previous codebook-based methods that rely on one-shot decoding, our method utilizes high-quality codes obtained in the previous iteration to guide the prediction of the Code-Predictor in the subsequent iteration, improving code prediction accuracy and ensuring stable dehazing performance. Our idea stems from the observations that 1) the degradation of hazy images varies with haze density and scene depth, and 2) clear regions play crucial cues in restoring dense haze regions. However, it is non-trivial to progressively refine the obtained codes in subsequent iterations, owing to the difficulty in determining which codes should be retained or replaced at each iteration. Another key insight of our study is to propose Code-Critic to capture interrelations among codes. The Code-Critic is used to evaluate code correlations and then resample a set of codes with the highest mask scores, i.e., a higher score indicates that the code is more likely to be rejected, which helps retain more accurate codes and predict difficult ones. Extensive experiments demonstrate the superiority of our method over state-of-the-art methods in real-world dehazing.

Problem

• Develops iterative code decoding for image dehazing.
• Improves code prediction accuracy using previous iterations.
• Introduces Code-Critic to evaluate and refine code correlations.

Innovation

• Iterative Predictor-Critic Code Decoding framework
• Leverages pre-trained VQGAN codebook prior
• Code-Critic evaluates and resamples high-quality codes