Selective Kernel Networks

Li, Xiang; Wang, Wenhai; Hu, Xiaolin; Yang, Jian

doi:10.1109/cvpr.2019.00060

Cited by 2,034 publications

(1,017 citation statements)

References 56 publications

Supporting

Mentioning

870

Contrasting

Order By: Relevance

“…In order to choose size-different kernel for each channel within concatenated multi-scale results, inspired by [13], a kernel selecting module is introduced. Details of a kernel selecting module is shown in Fig.…”

Section: Feature Fusion Stagementioning

confidence: 99%

Pyramid Real Image Denoising Network

Zhao

Jiang

Men

et al. 2019

2019 IEEE Visual Communications and Image Processing (VCIP)

View full text Add to dashboard Cite

While deep Convolutional Neural Networks (CNNs) have shown extraordinary capability of modelling specific noise and denoising, they still perform poorly on real-world noisy images. The main reason is that the real-world noise is more sophisticated and diverse. To tackle the issue of blind denoising, in this paper, we propose a novel pyramid real image denoising network (PRIDNet), which contains three stages. First, the noise estimation stage uses channel attention mechanism to recalibrate the channel importance of input noise. Second, at the multi-scale denoising stage, pyramid pooling is utilized to extract multi-scale features. Third, the stage of feature fusion adopts a kernel selecting operation to adaptively fuse multi-scale features. Experiments on two datasets of real noisy photographs demonstrate that our approach can achieve competitive performance in comparison with state-of-the-art denoisers in terms of both quantitative measure and visual perception quality. Code is available at https://github.com/491506870/PRIDNet.

show abstract

Section: Feature Fusion Stagementioning

confidence: 99%

Pyramid Real Image Denoising Network

Zhao

Jiang

Men

et al. 2019

2019 IEEE Visual Communications and Image Processing (VCIP)

View full text Add to dashboard Cite

show abstract

“…In this work, inspired by [28], we elaborately design an adaptive aggregation subnetwork to predict the modality weights in an end-to-end trained network. Note that our motivation is significantly different from [28] from the following can be found in Fig. 3 two aspects.…”

Section: Adaptive Aggregation Subnetworkmentioning

confidence: 99%

“…3 two aspects. First, [28] presents a nonlinear approach to aggregate information from multiple kernels to realize the adaptive receptive field sizes of neurons. However, we are trying to adaptively aggregate the information of different modalities, and do not consider the different receptive fields of the same feature layer.…”

Section: Adaptive Aggregation Subnetworkmentioning

confidence: 99%

Quality-Aware Feature Aggregation Network for Robust RGBT Tracking

Zhu

Tang

et al. 2021

IEEE Trans. Intell. Veh.

115

View full text Add to dashboard Cite

This paper investigates how to perform robust visual tracking in adverse and challenging conditions using complementary visual and thermal infrared data (RGBT tracking). We propose a novel deep network architecture called qualityaware Feature Aggregation Network (FANet) for robust RGBT tracking. Unlike existing RGBT trackers, our FANet aggregates hierarchical deep features within each modality to handle the challenge of significant appearance changes caused by deformation, low illumination, background clutter and occlusion. In particular, we employ the operations of max pooling to transform these hierarchical and multi-resolution features into uniform space with the same resolution, and use 1×1 convolution operation to compress feature dimensions to achieve more effective hierarchical feature aggregation. To model the interactions between RGB and thermal modalities, we elaborately design an adaptive aggregation subnetwork to integrate features from different modalities based on their reliabilities and thus are able to alleviate noise effects introduced by low-quality sources. The whole FANet is trained in an end-to-end manner. Extensive experiments on large-scale benchmark datasets demonstrate the high-accurate performance against other state-of-the-art RGBT tracking methods.

show abstract

“…This paper achieves this segmentation based on a modified U-net architecture. The squeeze-and-excitation residual (SE-Res) [15] and selective kernel (SK) modules [16] are respectively inserted in the down-sampling and up-sampling stages of the conventional U-net architecture. The SE-Res module considers more channel dependencies and lacks the spatial information of feature maps.…”

Section: Introductionmentioning

confidence: 99%

“…The spatial information is important for the pixel-level localization in the image segmentation task. The SK module is utilized to relieve this problem, which adaptively adjusts the size of local respective field in the convolutional operation to collect multi-scale spatial information [16]. The proposed SK-Unet framework has achieved robust segmentation performance in the LGE CMR sequences.…”

Section: Introductionmentioning

confidence: 99%

SK-Unet: An Improved U-Net Model with Selective Kernel for the Segmentation of Multi-sequence Cardiac MR

Wang

Yang

Tang

et al. 2020

Lecture Notes in Computer Science

View full text Add to dashboard Cite

In the clinical environment, myocardial infarction (MI) as one common cardiovascular disease is mainly evaluated based on the late gadolinium enhancement (LGE) cardiac magnetic resonance images (CMRIs). The automatic segmentations of left ventricle (LV), right ventricle (RV), and left ventricular myocardium (LVM) in the LGE CMRIs are desired for the aided diagnosis in clinic. To accomplish this segmentation task, this paper proposes a modified U-net architecture by combining multi-sequence CMRIs, including the cine, LGE, and T2-weighted CMRIs. The cine and T2-weighted CMRIs are used to assist the segmentation in the LGE CMRIs. In this segmentation network, the squeeze-and-excitation residual (SE-Res) and selective kernel (SK) modules are inserted in the down-sampling and up-sampling stages, respectively. The SK module makes the obtained feature maps more informative in both spatial and channel-wise space, and attains more precise segmentation result. The utilized dataset is from the MICCAI challenge (MS-CMRSeg 2019), which is acquired from 45 patients including three CMR sequences. The cine and T2weighted CMRIs acquired from 35 patients and the LGE CMRIs acquired from 5 patients are labeled. Our method achieves the mean dice score of 0.922 (LV), 0.827 (LVM), and 0.874 (RV) in the LGE CMRIs.

show abstract

Selective Kernel Networks

Cited by 2,034 publications

References 56 publications

Pyramid Real Image Denoising Network

Pyramid Real Image Denoising Network

Quality-Aware Feature Aggregation Network for Robust RGBT Tracking

SK-Unet: An Improved U-Net Model with Selective Kernel for the Segmentation of Multi-sequence Cardiac MR

Contact Info

Product

Resources

About