Contrastive Learning for Compact Single Image Dehazing

Wu, Haiyan; Qu, Yanyun; Lin, Shaohui; Zhou, Jian; Qiao, Ruizhi; Zhang, Zhizhong; Xie, Yuan; Ma, Lizhuang

doi:10.1109/cvpr46437.2021.01041

Cited by 433 publications

(192 citation statements)

References 30 publications

Supporting

Mentioning

190

Contrasting

Order By: Relevance

“…One downside of these approaches is that they are not designed to deal with natural noise in the real world. Wu et al [26] propose a single image dehazing method based on contrastive learning and achieves stateof-the-art performance on both synthetic and real world datasets but this method is trained in a supervised manner. This is not the case for some medical image modalities where ground truth is hard or impossible to acquire.…”

Section: General Denoising Methodsmentioning

confidence: 99%

Self-Supervised Bulk Motion Artifact Removal in Optical Coherence Tomography Angiography

Ren¹,

Park²,

Pan³

et al. 2022

Preprint

View full text Add to dashboard Cite

Optical coherence tomography angiography (OCTA) is an important imaging modality in many bioengineering tasks. The image quality of OCTA, however, is often hurt by Bulk Motion Artifacts (BMA), which are due to micromotion of subjects and typically appear as bright stripes surrounded by blurred areas. State-of-the-art BMA handling solutions usually treat the problem as an image inpainting one with deep neural network algorithms. These solutions, however, require numerous training samples with nontrivial annotation. Nevertheless, this context-based inpainting model has limited correction capability because it discards the rich structural and appearance information carried in the BMA stripe region. To address these issues, in this paper we propose a self-supervised content-aware BMA recover model. First, the gradient-based structural information and appearance feature are extracted from the BMA area and injected into the model to capture more connectivity. Second, with easily collected defective masks, the model is trained in a self-supervised manner that only the clear areas are for training while the BMA areas for inference. With structural information and appearance feature from noisy image as references, our model could correct larger BMA and produce better visualizing result. Only 2D images with defective masks are involved so our method is more efficient. Experiments on OCTA of mouse cortex demonstrate that our model could correct most BMA with extremely large sizes and inconsistent intensities while existing methods fail. Figure 1. Illustration of OCTA affected by Bulk Motion Artifacts (BMA). Our work focuses on the BMA removal task in vessel mask (blue rectangle in dashed line). The enhanced OCTA after BMA removal has better visualization quality.

show abstract

Section: General Denoising Methodsmentioning

confidence: 99%

Self-Supervised Bulk Motion Artifact Removal in Optical Coherence Tomography Angiography

Ren¹,

Park²,

Pan³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…In [ 22 ], a Cycle-Dehaze end-to-end network is proposed, the network does not require pairs of hazy/clean images to train the network Shao et al [ 26 ] proposed a domain adaptation framework called DA-dahazing, which consists of an image transformation module and two image dehazing modules, which first transforms the input hazy image from one domain to another, and then takes the transformed image and the original image as input to dehaze the image. Wu et al [ 27 ] proposed novel contrast regularization (CR) technology based on contrast learning that uses the information of the hazy image and the clear image as negative and positive samples, respectively.…”

Section: Related Workmentioning

confidence: 99%

High-Resolution Representations Network for Single Image Dehazing

Han

Zhu

et al. 2022

Sensors

View full text Add to dashboard Cite

Deep learning-based image dehazing methods have made great progress, but there are still many problems such as inaccurate model parameter estimation and preserving spatial information in the U-Net-based architecture. To address these problems, we propose an image dehazing network based on the high-resolution network, called DeHRNet. The high-resolution network originally used for human pose estimation. In this paper, we make a simple yet effective modification to the network and apply it to image dehazing. We add a new stage to the original network to make it better for image dehazing. The newly added stage collects the feature map representations of all branches of the network by up-sampling to enhance the high-resolution representations instead of only taking the feature maps of the high-resolution branches, which makes the restored clean images more natural. The final experimental results show that DeHRNet achieves superior performance over existing dehazing methods in synthesized and natural hazy images.

show abstract

“…Two-branch [33] utilized ensemble learning and transfer learning to enhance the fast learning and multi-learning capacity, and realized the model generalization performance in multi-class environments. AECR-Net [34] utilized novel contrast regularization (CR) technology based on contrast learning, hazy images and clear images were employed as negative samples and positive samples, respectively. To ensure the restored image was similar to the clear images, and away from the position of hazy images, CR was employed.…”

Section: Related Workmentioning

confidence: 99%

Lightweight multiple scale-patch dehazing network for real-world hazy image

2021

KSII TIIS

View full text Add to dashboard Cite

Image dehazing is an ill-posed problem which is far from being solved. Traditional image dehazing methods often yield mediocre effects and possess substandard processing speed, while modern deep learning methods perform best only in certain datasets. The haze removal effect when processed by said methods is unsatisfactory, meaning the generalization performance fails to meet the requirements. Concurrently, due to the limited processing speed, most dehazing algorithms cannot be employed in the industry. To alleviate said problems, a lightweight fast dehazing network based on a multiple scale-patch framework (MSP) is proposed in the present paper. Firstly, the multi-scale structure is employed as the backbone network and the multi-patch structure as the supplementary network. Dehazing through a single network causes problems, such as loss of object details and color in some image areas, the multipatch structure was employed for MSP as an information supplement. In the algorithm image processing module, the image is segmented up and down for processed separately. Secondly, MSP generates a clear dehazing effect and significant robustness when targeting real-world homogeneous and nonhomogeneous hazy maps and different datasets. Compared with existing dehazing methods, MSP demonstrated a fast inference speed and the feasibility of real-time processing. The overall size and model parameters of the entire dehazing model are 20.75M and 6.8M, and the processing time for the single image is 0.026s. Experiments on NTIRE 2018 and NTIRE 2020 demonstrate that MSP can achieve superior performance among the stateof-the-art methods, such as PSNR, SSIM, LPIPS, and individual subjective evaluation.

show abstract

Contrastive Learning for Compact Single Image Dehazing

Cited by 433 publications

References 30 publications

Self-Supervised Bulk Motion Artifact Removal in Optical Coherence Tomography Angiography

Self-Supervised Bulk Motion Artifact Removal in Optical Coherence Tomography Angiography

High-Resolution Representations Network for Single Image Dehazing

Lightweight multiple scale-patch dehazing network for real-world hazy image

Contact Info

Product

Resources

About