当前位置: X-MOL 学术J. Electron. Imaging › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
RGNAM: recurrent grid network with an attention mechanism for single-image dehazing
Journal of Electronic Imaging ( IF 1.0 ) Pub Date : 2021-06-01 , DOI: 10.1117/1.jei.30.3.033026
Nian Wang 1 , Zhigao Cui 1 , Yanzhao Su 1 , Aihua Li 1
Affiliation  

Single-image dehazing is a critical problem since haze existence degrades the quality of images and hinders most advanced computer vision tasks. Early methods solve this problem via the atmospheric scattering model, which estimate the intermediate parameters and then recover a clear image by low-level priors or learning on synthetic datasets. However, these model-based methods do not hold in various scenes. Recently, many learning-based methods have directly recovered dehazed images from inputs, but these methods fail to deal with dense haze and always lead to color distortion. To solve this problem, we build a recurrent grid network with an attention mechanism, named RGNAM. Specifically, we propose a recurrent feature extraction block, which repeats a local residual structure to enhance feature representation and adopts a spatial attention module to focus on dense haze. To alleviate color distortion, we extract local features (e.g., structures and edges) and global features (e.g., colors and textures) from a grid network and propose a feature fusion module combining trainable weights and channel attention mechanisms to merge these complementary features effectively. We train our model with smooth L1 loss and structural similarity loss. The experimental results demonstrate that our proposed RGNAM surpasses previous state-of-the-art single-image dehazing methods on both synthetic and real haze datasets.

中文翻译:

RGNAM:具有用于单图像去雾的注意机制的循环网格网络

单幅图像去雾是一个关键问题,因为雾霾的存在会降低图像质量并阻碍最先进的计算机视觉任务。早期的方法通过大气散射模型解决了这个问题,该模型估计中间参数,然后通过低级先验或对合成数据集的学习来恢复清晰的图像。然而,这些基于模型的方法并不适用于各种场景。最近,许多基于学习的方法直接从输入中恢复去雾图像,但这些方法无法处理密集的雾霾,并且总是导致颜色失真。为了解决这个问题,我们构建了一个带有注意力机制的循环网格网络,命名为 RGNAM。具体来说,我们提出了一个循环特征提取块,它重复局部残差结构以增强特征表示,并采用空间注意力模块专注于密集的雾霾。为了减轻颜色失真,我们从网格网络中提取局部特征(例如,结构和边缘)和全局特征(例如,颜色和纹理),并提出了一个结合可训练权重和通道注意机制的特征融合模块,以有效地合并这些互补特征。我们用平滑的 L1 损失和结构相似性损失来训练我们的模型。实验结果表明,我们提出的 RGNAM 在合成和真实雾度数据集上均优于先前最先进的单图像去雾方法。颜色和纹理),并提出了一个特征融合模块,结合了可训练的权重和通道注意机制,以有效地合并这些互补特征。我们用平滑的 L1 损失和结构相似性损失来训练我们的模型。实验结果表明,我们提出的 RGNAM 在合成和真实雾度数据集上均优于先前最先进的单图像去雾方法。颜色和纹理),并提出了一个特征融合模块,结合了可训练的权重和通道注意机制,以有效地合并这些互补特征。我们用平滑的 L1 损失和结构相似性损失来训练我们的模型。实验结果表明,我们提出的 RGNAM 在合成和真实雾度数据集上均优于先前最先进的单图像去雾方法。
更新日期:2021-06-17
down
wechat
bug