Can Large Pretrained Depth Estimation Models Help With Image Dehazing?
By: Hongfei Zhang , Kun Zhou , Ruizheng Wu and more
Potential Business Impact:
Clears foggy photos using smart color tricks.
Image dehazing remains a challenging problem due to the spatially varying nature of haze in real-world scenes. While existing methods have demonstrated the promise of large-scale pretrained models for image dehazing, their architecture-specific designs hinder adaptability across diverse scenarios with different accuracy and efficiency requirements. In this work, we systematically investigate the generalization capability of pretrained depth representations-learned from millions of diverse images-for image dehazing. Our empirical analysis reveals that the learned deep depth features maintain remarkable consistency across varying haze levels. Building on this insight, we propose a plug-and-play RGB-D fusion module that seamlessly integrates with diverse dehazing architectures. Extensive experiments across multiple benchmarks validate both the effectiveness and broad applicability of our approach.
Similar Papers
Can Large Pretrained Depth Estimation Models Help With Image Dehazing?
CV and Pattern Recognition
Clears foggy photos instantly using smart camera tricks.
Learning Implicit Neural Degradation Representation for Unpaired Image Dehazing
CV and Pattern Recognition
Clears foggy pictures without needing special rules.
Semi-supervised Image Dehazing via Expectation-Maximization and Bidirectional Brownian Bridge Diffusion Models
CV and Pattern Recognition
Clears foggy pictures, even in thick fog.