Lightweight Image Super-Resolution With Expectation-Maximization Attention Mechanism

Zhu, Xiangyuan; Guo, Ke; Ren, Siming; Hu, Bin; Hu, Min; Fang, Hui

doi:10.1109/tcsvt.2021.3078436

Cited by 109 publications

(38 citation statements)

References 57 publications

(74 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Owing to its strong optimization capability, the proposed GOMGBO can also be applied to other optimization problems, such as regression tasks, 141 medical diagnosis, [142][143][144][145] covert communication systems, [146][147][148] service ecosystem, 149,150 image editing, [151][152][153] energy storage planning and scheduling, 154 social recommendation and quality-of-service (QoS)-aware service composition, [155][156][157] active surveillance, 158 pedestrian dead reckoning, 159 evaluation of human lower limb motions, 160 image super resolution, [161][162][163] sentiment classification, 164 data-to-text generation, 165 crowd sensing, 166 and feature selection. [167][168][169]…”

Section: Experimental Results Of the Mammographic Data Setmentioning

confidence: 99%

Gaussian bare‐bones gradient‐based optimization: Towards mitigating the performance concerns

Qiao

Shan

Jiang

et al. 2021

Int J of Intelligent Sys

View full text Add to dashboard Cite

Gradient‐based optimizer (GBO) is a metaphor‐free mathematic‐based algorithm proposed in recent years. Encouraged by the gradient‐based Newton's method, this algorithm combines with population‐based evolutionary methods. The disadvantage of the traditional GBO algorithm is that the global search ability of the algorithm is too strong, and the local search ability is too weak; accordingly, it is difficult to obtain the global optimal solution efficiently. Therefore, a new improved GBO algorithm (GOMGBO) is developed to mitigate such performance concerns by introducing a Gaussian bare‐bones mechanism, an opposition‐based learning mechanism, and a moth spiral mechanism enhanced GBO algorithm. The proposed GOMGBO has been compared against many famous methods and improved variants on 30 benchmark functions. The experimental results show that GOMGBO has apparent advantages in convergence speed and precision. In addition, this paper analyzes the balance and diversity of the GOMGBO algorithm and compares GOMGBO with other algorithms on several engineering problems. The experimental results show that the GOMGBO algorithm is also better than the competitive algorithm in engineering problems. This study uses the GOMGBO algorithm to optimize kernel extreme learning machine (KELM), and a new GOMGBO‐KELM model is proposed. The model is used to deal with four clinical disease diagnosis problems. Compared with GBO‐KELM, back propagation neural network algorithm, and other models, comparative experiments show that GOMGBO‐KELM has high performance in dealing with practical cases. We invite the community to investigate further our method for solving problems more efficiently with reasonable speed and efficiency. Readers of this study can refer to https://aliasgharheidari.com for any guidance about the proposed GOMGBO method.

show abstract

Section: Experimental Results Of the Mammographic Data Setmentioning

confidence: 99%

Gaussian bare‐bones gradient‐based optimization: Towards mitigating the performance concerns

Qiao

Shan

Jiang

et al. 2021

Int J of Intelligent Sys

View full text Add to dashboard Cite

show abstract

“…Considering that our GAN-SR method currently performs only in-plane 2D image SR, to avoid the effect of slice thickness variability 6 , 7 , GAN-based SR along the z -axis (i.e., yielding thinner slices) might relieve the problem related to highly anisotropic voxels 50 , 51 . Moreover, since our GAN-SR model does not remarkably improve PSNR/SSIM values, we could conduct feature recalibration, such as via self-attention mechanisms, to obtain features more similar to the ones of the original images 21 , 52 – 54 . Concerning future radiomics applications, since we showed the results on a homogeneous subset of the NSCLC-Radiomics dataset, we plan to test the generalization ability of GAN-extracted radiomic features on the whole dataset, considering variations on CT image acquisition and reconstruction parameters.…”

Section: Discussionmentioning

confidence: 99%

“…To test the effectiveness of our framework, we compared it with other state-of-the-art methods, namely: Image Super-Resolution Network with an Expectation-Maximization Attention Mechanism (EMASRN 21 ), Enhanced Deep Super-Resolution (EDSR 29 ), Cascading Residual Network (CARN 30 ) and Super-Resolution based on Dictionary Learning and Sparse Representation (DLSR 31 ). For the EMASRN model, we relied on the implementation available at https://github.com/xyzhu1/EMASRN , optimizing the network for -norm loss during 1000 epochs with , a batch size of 16, and a learning rate of halved every 200 epochs.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Impact of GAN-based lesion-focused medical image super-resolution on the robustness of radiomic features

Farias

Noia

Han

et al. 2021

Sci Rep

View full text Add to dashboard Cite

Robust machine learning models based on radiomic features might allow for accurate diagnosis, prognosis, and medical decision-making. Unfortunately, the lack of standardized radiomic feature extraction has hampered their clinical use. Since the radiomic features tend to be affected by low voxel statistics in regions of interest, increasing the sample size would improve their robustness in clinical studies. Therefore, we propose a Generative Adversarial Network (GAN)-based lesion-focused framework for Computed Tomography (CT) image Super-Resolution (SR); for the lesion (i.e., cancer) patch-focused training, we incorporate Spatial Pyramid Pooling (SPP) into GAN-Constrained by the Identical, Residual, and Cycle Learning Ensemble (GAN-CIRCLE). At $$2\times $$ 2 × SR, the proposed model achieved better perceptual quality with less blurring than the other considered state-of-the-art SR methods, while producing comparable results at $$4\times $$ 4 × SR. We also evaluated the robustness of our model’s radiomic feature in terms of quantization on a different lung cancer CT dataset using Principal Component Analysis (PCA). Intriguingly, the most important radiomic features in our PCA-based analysis were the most robust features extracted on the GAN-super-resolved images. These achievements pave the way for the application of GAN-based image Super-Resolution techniques for studies of radiomics for robust biomarker discovery.

show abstract

“…SR is an ill-posed inverse problem since each LR patch has to be mapped onto multiple HR patches. To solve this problem, researchers have proposed a multitude of learning-based methods to learn the mapping function from LR patches to their HR counterparts [2][3][4][5][6][7][8][9][10][11][12][13].…”

Section: Introductionmentioning

confidence: 99%

Feedback Pyramid Attention Networks for Single Image Super-Resolution

Wu¹,

Gui²,

Zhang³

et al. 2021

Preprint

View full text Add to dashboard Cite

Recently, convolutional neural network (CNN) based image super-resolution (SR) methods have achieved significant performance improvement. However, most CNN-based methods mainly focus on feed-forward architecture design and neglect to explore the feedback mechanism, which usually exists in the human visual system. In this paper, we propose feedback pyramid attention networks (FPAN) to fully exploit the mutual dependencies of features. Specifically, a novel feedback connection structure is developed to enhance low-level feature expression with high-level information. In our method, the output of each layer in the first stage is also used as the input of the corresponding layer in the next state to re-update the previous low-level filters. Moreover, we introduce a pyramid non-local structure to model global contextual information in different scales and improve the discriminative representation of the network. Extensive experimental results on various datasets demonstrate the superiority of our FPAN in comparison with the state-of-the-art SR methods.

show abstract

Lightweight Image Super-Resolution With Expectation-Maximization Attention Mechanism

Cited by 109 publications

References 57 publications

Gaussian bare‐bones gradient‐based optimization: Towards mitigating the performance concerns

Gaussian bare‐bones gradient‐based optimization: Towards mitigating the performance concerns

Impact of GAN-based lesion-focused medical image super-resolution on the robustness of radiomic features

Feedback Pyramid Attention Networks for Single Image Super-Resolution

Contact Info

Product

Resources

About