Attention Optimized Deep Generative Adversarial Network for Removing Uneven Dense Haze

Zhao, Wenxuan; Zhao, Yaqin; Feng, Liqi; Tang, Jiaxi

doi:10.3390/sym14010001

Cited by 10 publications

(3 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The OSVOS is a typical algorithm framework based on independent segmentation. It does not consider the timing relationship and processes each frame independently, preventing the information of the previous and subsequent frames from interfering with the current frame [ 15 , 16 ]. The specific structure is given in Figure 2 .…”

Section: Relevant Theoretical Basis and Experimental Designmentioning

confidence: 99%

Film and Video Quality Optimization Using Attention Mechanism-Embedded Lightweight Neural Network Model

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

In filming, the collected video may be blurred due to camera shake and object movement, causing the target edge to be unclear or deforming the targets. In order to solve these problems and deeply optimize the quality of movie videos, this work proposes a video deblurring (VD) algorithm based on neural network (NN) model and attention mechanism (AM). Based on the scale recurrent network, Haar planar wavelet transform (WT) is introduced to preprocess the video image and to deblur the video image in the wavelet domain. Additionally, the spatial and channel AMs are fused into the overall network framework to improve the feature expression ability. Further, the residual inception spatial-channel attention (RISCA) mechanism is introduced to extract the multiscale feature information from video images. Meanwhile, skip spatial-channel attention (SSCA) accelerates the network training time to achieve a better VD effect. Finally, relevant experiments are designed, factoring in peak signal-to-noise ratio (PSNR) and structural similarity (SSI). The experimental findings corroborate that the proposed Haar and attention video deblurring (HAVD) outperforms multisize network Haar (MSNH) in PSNR and structural similarity (SSIM), improved by 0.10 dB and 0.005, respectively. Therefore, embedding the dual AMs can improve the model performance and optimize the video quality. This work provides technical support for solving the video distortion problems.

show abstract

Section: Relevant Theoretical Basis and Experimental Designmentioning

confidence: 99%

Film and Video Quality Optimization Using Attention Mechanism-Embedded Lightweight Neural Network Model

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…To monitor the quality of the generated images better, Peak Signal to Noise Ratio (PSNR) is also used as one of the evaluation metrics in dB, the larger the better [28]. In the calculation process of PSNR, two pictures need to be given, namely the predicted picture and the real picture, and a series of calculations are performed on the two to obtain the similarity score.…”

Section: A Implementation Detailsmentioning

confidence: 99%

CloudpredNet: An Ultra-Short-Term Movement Prediction Model for Ground-Based Cloud Image

Wei,

Zhu,

Guo

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Ground-based cloud images can provide information on weather and cloud conditions, which are important for cloud monitoring and PV power generation forecasting. Prediction of short-time cloud movement from images is a major means of intra-hourly irradiation forecasting for solar power generation and is also important for precipitation forecasting. However, there is a lack of advanced and complete methods for cloud movement prediction from ground-based cloud images, and traditional techniques based on image processing and motion vector calculations have difficulty in predicting cloud morphological changes, which makes accurate prediction of cloud motion (especially nonlinear motion) very challenging. Therefore, this paper proposes CloudpredNet, a ground-based cloud ultra-short-term movement prediction model based on an "encoder-generator" architecture. This paper also proposes a loss function dedicated to the time series prediction of ground-based cloud images and combines the attention mechanism to train the model. The model is validated on a publicly available dataset, and it is demonstrated that it has good performance in all metrics of cloud image generation for the next 10 minutes. Spatio-temporal prediction, ultra-short-term prediction, Transformer, ground-based cloud images INDEX TERMS

show abstract

“…Captured images in foggy weather have reduced contrast and brightness, which adversely causes difficulty for further perception and understanding for subsequent tasks. Therefore, haze removal, especially single image dehazing, is highly practical and realistic with comprehensive academic and industry value [1][2][3][4]. At present, researchers adopt a well-received physical model [5], which is formulated as as:…”

Section: Introductionmentioning

confidence: 99%

SNPD: Semi-Supervised Neural Process Dehazing Network with Asymmetry Pseudo Labels

et al. 2022

View full text Add to dashboard Cite

Haze can cause a significant reduction in the contrast and brightness of images. CNN-based methods have achieved benign performance on synthetic data. However, they show weak generalization performance on real data because they are only trained on fully labeled data, ignoring the role of natural data in the network. That is, there exists distribution shift. In addition to using little real data for training image dehazing networks in the literature, few studies have designed losses to constrain the intermediate latent space and the output simultaneously. This paper presents a semi-supervised neural process dehazing network with asymmetry pseudo labels. First, we use labeled data to train a backbone network and save intermediate latent features and parameters. Then, in the latent space, the neural process maps the latent features of real data to the latent space of synthetic data to generate one pseudo label. One neural process loss is proposed here. For situations where the image may be darker after dehazing, another pseudo label is created, and one new loss is used to guide the dehazing result at the output end. We combine the two pseudo labels with designed losses to suppress the distribution shift and guide better dehazing results. Finally, the artificial and hazy natural images are tested experimentally to demonstrate the method’s effectiveness.

show abstract

Attention Optimized Deep Generative Adversarial Network for Removing Uneven Dense Haze

Cited by 10 publications

References 52 publications

Film and Video Quality Optimization Using Attention Mechanism-Embedded Lightweight Neural Network Model

Film and Video Quality Optimization Using Attention Mechanism-Embedded Lightweight Neural Network Model

CloudpredNet: An Ultra-Short-Term Movement Prediction Model for Ground-Based Cloud Image

SNPD: Semi-Supervised Neural Process Dehazing Network with Asymmetry Pseudo Labels

Contact Info

Product

Resources

About