Subtopic Deep Dive
Single Image Dehazing
Research Guide
What is Single Image Dehazing?
Single Image Dehazing develops algorithms to remove haze from a single image using priors like color attenuation prior or deep neural networks based on atmospheric scattering models.
This subtopic addresses the ill-posed problem of haze removal from one image without depth information. Key methods include the color attenuation prior (Zhu et al., 2015, 2285 citations) and multi-scale CNNs (Ren et al., 2016, 1728 citations). Over 10 highly cited papers since 2015 advance deep learning approaches like FFA-Net (Xu et al., 2020, 1501 citations).
Why It Matters
Single image dehazing improves visibility in outdoor vision tasks like autonomous driving and surveillance under hazy weather. Zhu et al. (2015) enable fast processing for real-time applications in remote sensing. Ren et al. (2016) and Xu et al. (2020) support robust performance on real-world scenes, enhancing object detection accuracy in adverse conditions as shown in evaluations on hazy datasets.
Key Research Challenges
Handling Real-World Haze Variations
Algorithms struggle with diverse haze densities and lighting in real scenes beyond synthetic data. Zhu et al. (2015) use priors but falter on heavy haze. Xu et al. (2020) improve with attention but color distortions persist in evaluations.
Preserving Image Details
Dehazing often over-smooths textures or introduces halos near edges. Ren et al. (2016) apply multi-scale CNNs to retain details but require balanced scales. FFA-Net (Xu et al., 2020) fuses features yet challenges remain in fine structures.
Real-Time Computational Efficiency
Deep models like those in Wang et al. (2022) demand high compute unsuitable for edge devices. Zhu et al. (2015) offer speed but lack generalization. Balancing accuracy and speed persists as a core issue.
Essential Papers
A Fast Single Image Haze Removal Algorithm Using Color Attenuation Prior
Qingsong Zhu, Jiaming Mai, Ling Shao · 2015 · IEEE Transactions on Image Processing · 2.3K citations
Single image haze removal has been a challenging problem due to its ill-posed nature. In this paper, we propose a simple but powerful color attenuation prior for haze removal from a single input ha...
PVT v2: Improved baselines with pyramid vision transformer
Wenhai Wang, Enze Xie, Xiang Li et al. · 2022 · Computational Visual Media · 1.9K citations
Transformer recently has presented encouraging progress in computer vision.\nIn this work, we present new baselines by improving the original Pyramid Vision\nTransformer (PVT v1) by adding three de...
Single Image Dehazing via Multi-scale Convolutional Neural Networks
Wenqi Ren, Si Liu, Hua Zhang et al. · 2016 · Lecture notes in computer science · 1.7K citations
U2Fusion: A Unified Unsupervised Image Fusion Network
Han Xu, Jiayi Ma, Junjun Jiang et al. · 2020 · IEEE Transactions on Pattern Analysis and Machine Intelligence · 1.6K citations
This study proposes a novel unified and unsupervised end-to-end image fusion network, termed as U2Fusion, which is capable of solving different fusion problems, including multi-modal, multi-exposur...
FFA-Net: Feature Fusion Attention Network for Single Image Dehazing
Qin Xu, Zhilin Wang, Yuanchao Bai et al. · 2020 · Proceedings of the AAAI Conference on Artificial Intelligence · 1.5K citations
In this paper, we propose an end-to-end feature fusion at-tention network (FFA-Net) to directly restore the haze-free image. The FFA-Net architecture consists of three key components:1) A novel Fea...
Focal Loss for Dense Object Detection
Tsung-Yi Lin, Priya Goyal, Ross Girshick et al. · 2017 · arXiv (Cornell University) · 1.3K citations
The highest accuracy object detectors to date are based on a two-stage approach popularized by R-CNN, where a classifier is applied to a sparse set of candidate object locations. In contrast, one-s...
Deep Joint Rain Detection and Removal from a Single Image
Wenhan Yang, Robby T. Tan, Jiashi Feng et al. · 2017 · 1.2K citations
In this paper, we address a rain removal problem from a single image, even in the presence of heavy rain and rain streak accumulation. Our core ideas lie in our new rain image model and new deep le...
Reading Guide
Foundational Papers
Start with Zhu et al. (2015) for color attenuation prior as the simple, highly cited baseline explaining atmospheric modeling. Follow with Ren et al. (2016) to understand shift to multi-scale CNNs.
Recent Advances
Study Xu et al. (2020) FFA-Net for feature attention advances and Wang et al. (2022) PVT v2 for transformer baselines in enhancement tasks.
Core Methods
Core techniques: atmospheric scattering model inversion, dark channel/color attenuation priors (pre-2016), multi-scale CNNs, feature fusion attention (post-2016), pyramid vision transformers.
How PapersFlow Helps You Research Single Image Dehazing
Discover & Search
Research Agent uses searchPapers and citationGraph to map high-cite works like Zhu et al. (2015) with 2285 citations, then findSimilarPapers reveals FFA-Net (Xu et al., 2020) and Ren et al. (2016); exaSearch uncovers niche priors in 250M+ OpenAlex papers.
Analyze & Verify
Analysis Agent applies readPaperContent to extract atmospheric models from Zhu et al. (2015), verifies claims with CoVe on haze removal metrics, and runs PythonAnalysis to compute PSNR/SSIM stats on O-HAZE dataset; GRADE scores evidence strength for priors vs. CNNs.
Synthesize & Write
Synthesis Agent detects gaps like real-world generalization post-Xu et al. (2020), flags contradictions in priors; Writing Agent uses latexEditText, latexSyncCitations for Zhu/Ren papers, latexCompile for reports, exportMermaid for network architecture diagrams.
Use Cases
"Reimplement color attenuation prior from Zhu 2015 in Python and test on hazy images"
Research Agent → searchPapers('Zhu 2015 haze') → Analysis Agent → readPaperContent + runPythonAnalysis (NumPy PSNR calc) → researcher gets executable code + metrics plot.
"Compare FFA-Net vs multi-scale CNN dehazing in LaTeX survey table"
Research Agent → citationGraph → Synthesis → gap detection → Writing Agent → latexEditText + latexSyncCitations (Xu 2020, Ren 2016) + latexCompile → researcher gets PDF with comparison table and diagrams.
"Find GitHub repos for single image dehazing models like FFA-Net"
Research Agent → searchPapers('FFA-Net Xu 2020') → Code Discovery → paperExtractUrls → paperFindGithubRepo → githubRepoInspect → researcher gets repo code, benchmarks, and setup instructions.
Automated Workflows
Deep Research workflow scans 50+ dehazing papers via searchPapers → citationGraph → structured report on priors to CNNs evolution (Zhu 2015 to Wang 2022). DeepScan applies 7-step CoVe verification on Ren et al. (2016) methods with GRADE checkpoints and Python metric runs. Theorizer generates hypotheses on hybrid prior+CNN models from literature gaps.
Frequently Asked Questions
What defines single image dehazing?
Single image dehazing removes haze from one RGB image using priors or networks without multi-view or depth data, based on atmospheric scattering models.
What are key methods in single image dehazing?
Early methods use color attenuation prior (Zhu et al., 2015). Modern approaches employ CNNs (Ren et al., 2016) and attention networks like FFA-Net (Xu et al., 2020).
What are the most cited papers?
Top papers include Zhu et al. (2015, 2285 citations), Ren et al. (2016, 1728 citations), and Xu et al. (2020, 1501 citations).
What open problems exist?
Challenges include real-world haze variability, detail preservation, and real-time efficiency on edge devices, as noted in evaluations beyond synthetic datasets.
Research Image Enhancement Techniques with AI
PapersFlow provides specialized AI tools for Computer Science researchers. Here are the most relevant for this topic:
AI Literature Review
Automate paper discovery and synthesis across 474M+ papers
Code & Data Discovery
Find datasets, code repositories, and computational tools
Deep Research Reports
Multi-source evidence synthesis with counter-evidence
AI Academic Writing
Write research papers with AI assistance and LaTeX support
See how researchers in Computer Science & AI use PapersFlow
Field-specific workflows, example queries, and use cases.
Start Researching Single Image Dehazing with AI
Search 474M+ papers, run AI-powered literature reviews, and write with integrated citations — all in one workspace.
See how PapersFlow works for Computer Science researchers
Part of the Image Enhancement Techniques Research Guide