Progressive Image Deraining Networks: A Better and Simpler Baseline

Ren, Dongwei; Zuo, Wangmeng; Hu, Qinghua; Zhu, Pengfei; Meng, Deyu

doi:10.1109/cvpr.2019.00406

Cited by 733 publications

(468 citation statements)

References 35 publications

Supporting

Mentioning

464

Contrasting

Unclassified

Order By: Relevance

“…The DID-MDN [6] model may remove the objects which context is similar to rain streaks due to losing distribution of rain streaks. PReNet [16] can achieve good performance on synthetic dataset. However, its performance is not satisfying in real-world images.…”

Section: Results On Real-world Imagesmentioning

confidence: 96%

“…Yang et al [8] propose a multi-task network called Joint Rain Detection and Removal (JORDER) to learn the binary rain streak map, the appearance of rain streaks, and the clean background. Ren et al [16] propose a network called progressive recurrent network (PReNet). The PReNet takes advantage of recursive computation by repeatedly unfolding a shallow ResNet.…”

Section: Cnn-based Methodsmentioning

confidence: 99%

“…Thus, the proposed model outperforms other GAN-based models. The proposed model is also compared with other state-of-the-art deep learning-based methods as follows: JORDER [8], DDN [7], DID-MDN [6] and PReNet [16]. Quantitative results are tabulated in Table 4.…”

Section: Comparison With State-of-the-art Methodsmentioning

confidence: 99%

“…Visual comparison results on synthesic DID-MDN dataset. From left to right: Input, JORDER [8], DDN [7], DID-MDN [6], PReNet [16], GAN-SID [13], Ours, Ground Truth.…”

Section: Comparison With State-of-the-art Methodsmentioning

confidence: 99%

“…DDN[7] (d) DID-MDN[6] (e) PReNet[16] (f) GAN-SID[13] (g) Ours Visual comparison results on real-world images.…”

unclassified

See 4 more Smart Citations

Single Image De-Raining via Improved Generative Adversarial Nets

Ren

Nie

et al. 2020

Sensors

View full text Add to dashboard Cite

Capturing images under rainy days degrades image visual quality and affects analysis tasks, such as object detection and classification. Therefore, image de-raining has attracted a lot of attention in recent years. In this paper, an improved generative adversarial network for single image de-raining is proposed. According to the principles of divide-and-conquer, we divide an image de-raining task into rain locating, rain removing, and detail refining sub-tasks. A multi-stream DenseNet, termed as Rain Estimation Network, is proposed to estimate the rain location map. A Generative Adversarial Network is proposed to remove the rain streaks. A Refinement Network is proposed to refine the details. These three models accomplish rain locating, rain removing, and detail refining sub-tasks, respectively. Experiments on two synthetic datasets and real world images demonstrate that the proposed method outperforms state-of-the-art de-raining studies in both objective and subjective measurements.

show abstract

Section: Results On Real-world Imagesmentioning

confidence: 96%

Section: Cnn-based Methodsmentioning

confidence: 99%

Section: Comparison With State-of-the-art Methodsmentioning

confidence: 99%

“…Visual comparison results on synthesic DID-MDN dataset. From left to right: Input, JORDER [8], DDN [7], DID-MDN [6], PReNet [16], GAN-SID [13], Ours, Ground Truth.…”

Section: Comparison With State-of-the-art Methodsmentioning

confidence: 99%

“…DDN[7] (d) DID-MDN[6] (e) PReNet[16] (f) GAN-SID[13] (g) Ours Visual comparison results on real-world images.…”

unclassified

See 3 more Smart Citations

Single Image De-Raining via Improved Generative Adversarial Nets

Ren

Nie

et al. 2020

Sensors

View full text Add to dashboard Cite

show abstract

Degradation of Commercially Available Digital Camera Images due to Variation of Rainfall Intensity in Outdoor Conditions

Kanazawa

Uchida

2022

Preprint

View full text Add to dashboard Cite

Camera-based rainfall observation is a useful technology that contributes to the densification of rainfall observation networks because it can measure rainfall with high spatio-temporal resolution and low cost. To develop of practical camera-based rainfall observation technology, using the extinction coefficient as a clue, this study proposed relational Equations representing the relationship between image information, rainfall intensity, and scene depth by linking the theoretically derived rainfall intensity with a technique proposed in the computer vision field for removing static weather effects. Then, the proposed Equations were applied to outdoor images taken by commercial interval cameras at the observation site in a mountainous watershed in Japan. As a result, it was confirmed that transmission calculated from the image information decreases exponentially according to the increase in rainfall intensity and scene depth, as assumed in the proposed Equations. Therefore, the proposed Equations are generally valid even for outdoor images, and extremely important findings that will improve camera-based rainfall observation techniques were obtained. On the other hand, the calculated extinction coefficient tended to be overestimated in patches with a small scene depth, and overestimation of the extinction coefficient due to aerosol effects was also observed in the images taken during no rainfall. Although there are issues at present that need to be resolved for the technology proposed in this study, this technology has the potential to help the development of a camera-based rainfall observation technology that is accurate, robust, versatile, and accessible.

show abstract

Single image deraining using multi‐stage and multi‐scale joint channel coordinate attention fusion network

Yang

Zhang

Cui

et al. 2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

Rain streaks can seriously degrade the visual quality of an image and are detrimental to subsequent algorithms such as object detection and semantic segmentation. Therefore, removing rain streaks is a very important task. The deraining task has two main limitations: the first is to encode information about rain streaks in different densities and directions, the second is to keep the background details of the image while removing the rain streak. To address these limitations, we propose an effective algorithm, called multi-stage and multi-scale joint channel coordinate attention fusion network (MMAFN). We mainly propose a two-stage network structure, both of which use an encoderdecoder network to extract features. The first-stage network extracts coarse features and the second-stage network integrates the features of the former to further refine features. We design the joint channel coordinate attention block to encode features of rain streaks in different directions and densities. In addition, to better fuse features of different scales and enhance the generalization performance of the network, the inception attention branch block and the multi-level feature fusion block are designed. Extensive experiments substantiate the superiority of the proposed network

show abstract

Progressive Image Deraining Networks: A Better and Simpler Baseline

Cited by 733 publications

References 35 publications

Single Image De-Raining via Improved Generative Adversarial Nets

Single Image De-Raining via Improved Generative Adversarial Nets

Degradation of Commercially Available Digital Camera Images due to Variation of Rainfall Intensity in Outdoor Conditions

Single image deraining using multi‐stage and multi‐scale joint channel coordinate attention fusion network

Contact Info

Product

Resources

About