PaperlayerView exact revision on arXiv

Source-linked AI summary

FAMED-Net: A Fast and Accurate Multi-scale End-to-end Dehazing Network

Jing Zhang, Dacheng Tao

arXiv:1906.04334v2cs.CVcs.LG

TL;DR

Single-image dehazing is an ill-posed restoration problem important for downstream vision tasks, while existing models may be inefficient or limited in representation capacity. FAMED-Net uses a lightweight three-scale, point-wise end-to-end architecture to learn haze-free images directly. Experiments report superiority over representative models across complexity, efficiency, restoration accuracy, and cross-set generalization.

  • Problem

    Single-image dehazing remains ill-posed, while existing models can be overcomplex, computationally inefficient, or limited in representation capacity.

    Source: §I · pp. 1–2

  • Method

    FAMED-Net combines three-scale encoders and a fusion module with densely connected point-wise convolutions and pooling layers to learn haze-free images directly.

    Source: §I · p. 2

  • Results

    FAMED-Net outperforms representative state-of-the-art models on synthetic benchmarks and real-world hazy images across model complexity, computational efficiency, restoration accuracy, and cross-set generalization.

    Source: §I · p. 2, §V · p. 12

  • Takeaways & Limitations

    FAMED-Net provides a fast, lightweight, and accurate deep architecture for single-image dehazing.

    Source: §I · p. 2, §V · p. 12

    Transmission estimates can be incorrect in sky regions, producing under-dehazed artifacts; incorporating high-level semantics is left for future work.

    Source: D. Limitations and Discussions · pp. 11–12

Abstract

from arXiv · show

Single image dehazing is a critical image pre-processing step for subsequent high-level computer vision tasks. However, it remains challenging due to its ill-posed nature. Existing dehazing models tend to suffer from model overcomplexity and computational inefficiency or have limited representation capacity. To tackle these challenges, here we propose a fast and accurate multi-scale end-to-end dehazing network called FAMED-Net, which comprises encoders at three scales and a fusion module to efficiently and directly learn the haze-free image. Each encoder consists of cascaded and densely connected point-wise convolutional layers and pooling layers. Since no larger convolutional kernels are used and features are reused layer-by-layer, FAMED-Net is lightweight and computationally efficient. Thorough empirical studies on public synthetic datasets (including RESIDE) and real-world hazy images demonstrate the superiority of FAMED-Net over other representative state-of-the-art models with respect to model complexity, computational efficiency, restoration accuracy, and cross-set generalization. The code will be made publicly available.

I. INTRODUCTION

Single-image dehazing is a challenging ill-posed restoration problem that matters for downstream vision tasks. FAMED-Net addresses this with a lightweight multi-scale end-to-end architecture and outperforms representative state-of-the-art models across synthetic and real-world images.

  • Haze causes low-contrast, blurry, and noisy images that can hinder object detection and semantic segmentation.Source: §I · p. 1
  • Single-image dehazing is practically useful because it requires no additional scene depth, polarization, or image sequence, but remains ill-posed.Source: §I · p. 1
  • FAMED-Net uses three-scale encoders and a fusion module to directly learn the haze-free image end-to-end.Source: §I · p. 2
  • Fully point-wise convolutions, densely reused features, and pooling layers make FAMED-Net lightweight and computationally efficient.Source: §I · p. 2
  • FAMED-Net outperforms representative state-of-the-art models on synthetic benchmarks and real-world hazy images.Source: §I · p. 2

II. RELATED WORK

Prior-based dehazing estimates haze structure using assumptions such as transmission, atmospheric light, and dark-channel statistics, while learning-based methods learn mappings from data. End-to-end models increasingly learn scene radiance directly rather than estimating transmission first.

C. Multi-scale pyramid architecture

The paper frames dehazing through multi-scale feature representation and a probabilistic formulation, using point-wise convolutions to learn statistical regularities efficiently.

Multi-scale representation

Deep supervision

  • Deep supervision adds intermediate dehazed-image supervision at each scale to facilitate multi-scale and multi-level feature learning.Source: D. Deep Supervision · p. 3

Probabilistic formulation

Point-wise modeling

B. The Single-scale FAMED-Net: FAMED-Net-SS

FAMED-Net-SS learns a latent K variable with a fully point-wise CNN, reusing features across receptive fields and aggregating multi-level statistics through pooling.

Network design

Multi-level aggregation

Training

C. The Multi-scale Variants of FAMED-Net: FAMED-Net-GP and FAMED-Net-LP

FAMED-Net extends the single-scale design with Gaussian and Laplacian pyramid variants that combine multi-scale K estimates before recovering scene radiance.

FAMED-Net-GP

Training

FAMED-Net-LP

D. Model Complexity Analysis

FAMED-Net is designed to remain lightweight and computationally efficient through fully point-wise convolutions, while supporting arbitrary image sizes and fixed-size testing.

Complexity

Scalability

Testing strategy

  • Fixed-size testing resizes the longest image side to 360, then restores the estimated K map to the original size by bilinear interpolation and refinement.Source: D. Model Complexity Analysis · p. 6

Comparison

  • Table II compares FAMED-Net and state-of-the-art models using parameters, model size, and runtime.Source: Table II · p. 6

IV. EXPERIMENTS

Experiments evaluate FAMED-Net against prior-based and deep CNN methods on RESIDE, using synthetic benchmark scores, real-world comparisons, and ablations. Training uses RESIDE’s large synthetic datasets and 128×128 patches, with implementation settings and architecture variants documented.

Evaluation setup

  • RESIDE evaluation compares FAMED-Net with image prior-based and deep CNN-based dehazing methods.The benchmark reports PSNR and SSIM on SOTS and includes subjective evaluation on real-world hazy images.Source: §IV · p. 6

Datasets

  • RESIDE provides 110,500 synthetic indoor and 313,950 synthetic outdoor hazy training images, while SOTS contains 500 indoor and 500 outdoor test images.Source: §IV · p. 6

Training protocol

  • FAMED-Net training uses 128×128 randomly cropped patches, with depth-level statistics analyzed across the training patches.Source: §IV · p. 6, Figure 4 · p. 6
  • FAMED-Net is trained for 400,000 iterations, using a batch size of 48 and a learning rate reduced after 200,000 and 320,000 iterations.Source: §IV · p. 6

Ablation design

1) Ablations on the Basic Architecture:

Ablations show that wider features, batch normalization, multi-scale processing, and sufficient training data improve FAMED-Net, while added structural convolutions improve accuracy at modest computational cost. The Gaussian-pyramid model is selected as the accuracy–complexity compromise.

Basic architecture

Feature channels

Multi-scale design

Training data and iterations

Pyramid variants

3) The Effectiveness of Max Pooling:

Max pooling substantially improves the multi-scale model while preserving its lightweight design, and FAMED-Net achieves the strongest reported SOTS comparison alongside favorable qualitative results. The architecture uses point-wise convolutions, multi-scale processing, and dense feature reuse.

Max pooling

Efficiency

Quantitative comparison

Design conclusions

  • The authors attribute the compact model to point-wise convolutions, layer-by-layer feature aggregation, end-to-end modeling, multi-scale design, and dense feature reuse.Source: 1) Results on RESIDE SOTS · p. 8

Subjective evaluation

  • On synthetic hazy images, competing methods exhibit residual haze, over-saturation, color distortions, or noise, whereas FAMED-Net produces comparatively cleaner results.Source: 2) Subjective Evaluation · p. 8, Figure 5 · p. 9
  • On real-world images, FAMED-Net produces better or comparable results with clear details and fewer color artifacts than representative state-of-the-art methods.Source: 2) Subjective Evaluation · pp. 8–9

3) Cross-set Generalization:

FAMED-Net is evaluated for cross-set generalization without fine-tuning and for learned statistical regularities on haze-free images. It generalizes better than GFN and DCPDN and avoids the over-dehazing tendency associated with AOD-Net’s learned statistics.

Cross-set generalization

Learned regularities

Transmission visualization

  • The transmission visualization uses warm colors for high transmission, corresponding to near-camera regions with small depth.Source: Figure 9 · p. 11

5) Runtime Analysis:

FAMED-Net is evaluated as a lightweight, efficient dehazing architecture, while the discussion identifies limitations in its learned statistical regularity and current handling of heterogeneous illumination and haze.

  • FAMED-Net is described as outperforming prior methods while using a more efficient architecture.The paper frames this advantage in the context of its limitations discussion.Source: D. Limitations and Discussions · p. 11
  • FAMED-Net implicitly learns locally statistical regularity for dehazing, similarly to prior- and learning-based methods.Source: D. Limitations and Discussions · p. 11
  • Transmission estimates are incorrect in sky regions, producing under-dehazed artifacts.The authors suggest incorporating high-level semantics and jointly modeling enhancement with high-level understanding as future work.Source: D. Limitations and Discussions · p. 12
  • The authors propose extending FAMED-Net to non-uniform atmospheric light and heterogeneous haze using illumination modeling and region-based techniques.Examples include artificial ambient light in nighttime haze and haze-density-aware segmentation.Source: D. Limitations and Discussions · p. 12
  • Experiments on benchmark datasets and real-world hazy images support FAMED-Net as a fast, lightweight, and accurate dehazing architecture.Source: §V · p. 12

A. Illumination Balancing Network

IBNet adapts FAMED-Net to estimate balanced illumination for dehazed images, with experiments reporting accurate restoration and visually improved illumination, detail, and comparisons against competing methods.

Motivation

  • Dehazed images can appear dim because scene radiance is usually less bright than atmospheric light, motivating illumination balancing.The stated goals are more pleasing appearance and support for subsequent high-level tasks.Source: 1) Modification of FAMED-Net for… · p. 14

Network design

  • IBNet estimates a balanced illumination map and replaces the input image’s unbalanced illumination distribution with that estimate.It is constructed from FAMED-Net with a one-channel output and without the recovery module.Source: 1) Modification of FAMED-Net for… · p. 14

Dataset preparation

Results

  • IBNet achieved good restoration accuracy on RESIDE TestSet-S for illumination balancing.The supplied passage references PSNR and SSIM results in Table VIII but does not provide their values.Source: Table VIII · p. 14

Comparisons

Related papers

Loading 1906.04334v2…