Atmospheric Turbulence Correction via Variational Deep Diffusion

Atmospheric Turbulence (AT) correction is a challenging restoration task as it consists of two distortions: geometric distortion and spatially variant blur. If left unaddressed, these distortions can significantly impact downstream vision tasks. Diffusion models have shown impressive accomplishments in photo-realistic image synthesis and beyond. In this paper, we propose a novel deep conditional diffusion model under a variational inference framework to solve the AT correction problem. We use this framework to improve performance by learning latent prior information from the input and degradation processes. We use the learned information to further condition the diffusion model. Experiments are conducted in a comprehensive synthetic AT dataset. We show that the proposed framework achieves good quantitative and qualitative results.

• Publication year

  2023

• Venue

  Conference on Multimedia Information Processing and Retrieval

• Publication date

  2023-05-08

• Fields of study

  Computer Science, Engineering, Environmental Science

• Identifiers
  DOI 10.1109/MIPR59079.2023.00022arXiv 2305.05077
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Variational Deep Image Restoration

  2022influential reference
• Removing Atmospheric Turbulence via Deep Adversarial Learning

  2022cited by this paper
• AT-DDPM: Restoring Faces Degraded by Atmospheric Turbulence Using Denoising Diffusion Probabilistic Models

  2022cited by this paper
• Single Frame Atmospheric Turbulence Mitigation: A Benchmark Study and A New Physics-Inspired Transformer Model

  2022cited by this paper
• Accelerating Atmospheric Turbulence Simulation via Learned Phase-to-Space Transform

  2021cited by this paper
• Generative Adversarial Networks

  2021cited by this paper
• High-Resolution Image Synthesis with Latent Diffusion Models

  2021cited by this paper
• Image Super-Resolution via Iterative Refinement

  2021cited by this paper
• Deep learning for tomographic image reconstruction

  2020cited by this paper
• Denoising Diffusion Probabilistic Models

  2020cited by this paper
• NTIRE 2019 Challenge on Video Deblurring and Super-Resolution: Dataset and Study

  2019cited by this paper
• Spectral Normalization for Generative Adversarial Networks

  2018cited by this paper
• The Unreasonable Effectiveness of Deep Features as a Perceptual Metric

  2018cited by this paper
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium

  2017cited by this paper
• Wide Residual Networks

  2016cited by this paper
• SGDR: Stochastic Gradient Descent with Warm Restarts

  2016cited by this paper
• U-Net: Convolutional Networks for Biomedical Image Segmentation

  2015cited by this paper
• Adam: A Method for Stochastic Optimization

  2014cited by this paper
• Making a “Completely Blind” Image Quality Analyzer

  2013cited by this paper
• Removing Atmospheric Turbulence via Space-Invariant Deconvolution

  2013cited by this paper

• Tooth alignment by diffusion model with hierarchical spatial relationship learning

  2026cites this paper
• Learning Phase Distortion with Selective State Space Models for Video Turbulence Mitigation

  2025cites this paper
• Thermal turbulence mitigation in underwater images of nuclear fuel assemblies

  2025cites this paper
• Improving classification accuracy in post-turbulence-mitigation tasks via perceptual loss optimization

  2025cites this paper
• Industrial Anomaly Detection Based on Improved Diffusion Model: A Review

  2025cites this paper
• 大气湍流图像模拟及校正方法研究进展（特邀）

  2025cites this paper
• Diffusion Models, Image Super-Resolution, and Everything: A Survey

  2024cites this paper
• Deep learning-based turbulence mitigation for long range imaging

  2024cites this paper
• Real-World Atmospheric Turbulence Correction via Domain Adaptation

  2024cites this paper
• DeTurb: Atmospheric Turbulence Mitigation with Deformable 3D Convolutions and 3D Swin Transformers

  2024cites this paper
• A Comprehensive Survey on Diffusion Models and Their Applications

  2024influential citation
• Deep learning techniques for atmospheric turbulence removal: a review

  2024influential citation
• Removal of atmospheric turbulence-degraded effect using multi-scale features from images at different wavelengths

  2024cites this paper