Efficient and Accurate Multi-Scale Topological Network for Single Image Dehazing

Yi, Qiaosi; Li, Juncheng; Fang, Faming; Jiang, Aiwen; Zhang, Guixu

doi:10.1109/tmm.2021.3093724

Cited by 28 publications

(12 citation statements)

References 68 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ASFF assigns weights to features to suppress feature conflicts between different scales, thus improving the scale invariance of features. Yi et al [ 40 ] proposed an adaptive feature selection module (AFSM) for fusing neighboring scale features to improve the image dehazing effect. However, both ASFF and AFSM reduce the channel dimension so as to reduce the number of parameters, which leads to the underutilization of features.…”

Section: Methodsmentioning

confidence: 99%

“…AFAB solves the inconsistency of the pyramid and enables the deep semantic information to better guide the shallow fine-grained information to learn finer textures and colors. To illustrate the advantages of the proposed AFAB, we further compare it with ASFF [39] and AFSM [40]. Since AFSM only performs adaptive fusion of neighboring scale features, we improved it to receive features from each layer of the pyramid.…”

Section: Effectiveness Of Network Structurementioning

confidence: 99%

See 1 more Smart Citation

Efficient adaptive feature aggregation network for low-light image enhancement

Gao

Liu

et al. 2022

PLoS ONE

View full text Add to dashboard Cite

Existing learning-based methods for low-light image enhancement contain a large number of redundant features, the enhanced images lack detail and have strong noises. Some methods try to combine the pyramid structure to learn features from coarse to fine, but the inconsistency of the pyramid structure leads to luminance, color and texture deviations in the enhanced images. In addition, these methods are usually computationally complex and require high computational resource requirements. In this paper, we propose an efficient adaptive feature aggregation network (EAANet) for low-light image enhancement. Our model adopts a pyramid structure and includes multiple multi-scale feature aggregation block (MFAB) and one adaptive feature aggregation block (AFAB). MFAB is proposed to be embedded into each layer of the pyramid structure to fully extract features and reduce redundant features, while the AFAB is proposed for overcome the inconsistency of the pyramid structure. EAANet is very lightweight, with low device requirements and a quick running time. We conducted an extensive comparison with some state-of-the-art methods in terms of PSNR, SSIM, parameters, computations and running time on LOL and MIT5K datasets, and the experiments show that the proposed method has significant advantages in terms of comprehensive performance. The proposed method reconstructs images with richer color and texture, and the noises is effectively suppressed.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Effectiveness Of Network Structurementioning

confidence: 99%

Efficient adaptive feature aggregation network for low-light image enhancement

Gao

Liu

et al. 2022

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…The captured image under haze scenes can significantly affect image processing tasks such as object recognition and so on [4][5][6]. Therefore, image dehazing [7][8][9] has always been a hot research field in image processing.…”

Section: Introductionmentioning

confidence: 99%

Black-box Dehazing Network via Joint Loss Function

Zhang,

Zhao,

Song

et al. 2024

Preprint

View full text Add to dashboard Cite

Single image dehazing is a notoriously challenging task in image processing due to the numerous unknown factors involved. Most existing dehazing methods are based on physical models of the haze formation process. However, real-world haze scenes cannot be accurately mathematically modeled due to the existence of various unquantifiable factors in the scene. Therefore, the dehazing methods based on physical models often perform poorly in complex haze scenes. In this paper, we propose a novel black-box dehazing equation. In this equation, the haze is modelled as an additional image interference layer, without explicitly reasoning about the physical model of haze formation. The dehazing process is modelled as removing this image interference layer. Based on this equation, we propose a novel network architecture called the Black-box Dehazing Network (BDN). Moreover, we propose a joint loss function for training this network. The joint loss function not only evaluates pixel-level differences between the dehazed image and the haze-free image, but also measures differences in texture, color, and structure between the hazy image and its corresponding dehazed version as well as those between the hazy image and its haze-free version. In training, BDNet is only fed pairs of haze-free images and their corresponding hazy images. The corresponding hazy patches are generated on-the-fly during network training. Experimental results demonstrate that the proposed method has the advantage of universality and outperforms existing state-of-the-art dehazing methods.

show abstract

“…First, the coarse‐scale network is used to predict the overall projection image to obtain a rough dehazing image, and then the fine‐scale network uses high‐frequency detail information to repair and obtain a clear image. [20] proposed a multi‐scale topological network (MSTN) to explore features at different scales. At the same time, multi‐scale feature fusion module (MFFM) and adaptive feature selection module (AFSM) are designed to realize feature selection and fusion at different scales.…”

Section: Introductionmentioning

confidence: 99%

Two‐stage single image dehazing network using swin‐transformer

Hua

2022

IET Image Processing

View full text Add to dashboard Cite

Hazy images often have color distortion, blur and other visible visual quality degradation, affecting the performance of some advanced visual tasks. Therefore, single image dehazing has always been a challenging and significant problem. Convolutional neural network has been widely used in image dehazing task, but the limitations of convolutional operation limit the development of dehazing task. Nowadays, Transformer offers a holistic approach to CV development and does not grow in location as the network deepens. For this reason, a hierarchical Transformer is introduced for use in the dehazing network. Specifically, the codec is improved and Transformer and CNN are combined to achieve basic feature extraction in the first stage. The encoder only models the global relationship at each layer, reducing the resolution of the feature map continuously and expanding the field of perception. In addition, an inter‐block supervision mechanism is added between encoder unit and decoder unit to refine features and supervise and select them, thus improving the efficiency of feature transmission. In the second stage, the original resolution block is used to extract the local features, and then feature fusion and interaction are carried out. In addition, to ensure the authenticity of the transmission of characteristic signals in the first stage and improve the transmission efficiency of the network, fusion attention mechanism is added between stages. It adds the residual image of the early input features to the image acquired in the first stage, then passes to the next stage. Ablation experiments show that the two‐stage network has significant benefits for image quality and visual effects. The experimental results on RESIDE, O‐Haze, and I‐Haze datasets show that the method is superior to advanced methods in dehazing effectiveness.

show abstract

Efficient and Accurate Multi-Scale Topological Network for Single Image Dehazing

Cited by 28 publications

References 68 publications

Efficient adaptive feature aggregation network for low-light image enhancement

Efficient adaptive feature aggregation network for low-light image enhancement

Black-box Dehazing Network via Joint Loss Function

Two‐stage single image dehazing network using swin‐transformer

Contact Info

Product

Resources

About