Lightweight global-locally connected distillation network for single image super-resolution

Zeng, Cong; Li, Guangyao; Chen, Qiaochuan; Xiao, Qingguo

doi:10.1007/s10489-022-03454-y

Cited by 3 publications

(3 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We compare our IRAN with other state-of-the-art lightweight SR methods, including bicubic, SRCNN, 4 FSRCNN, 37 VDSR, 7 LapSRN, 18 DRRN, 12 MemNet, 38 CARN, 13 CARN-M, 13 IMDN, 14 LMRN, 15 MADNet, 16 MFEN-S, 17 GLCDNet, 19 ELCRN, 21 and WMRN 39 Tables 1–3. show the quantitative comparison results for different upscale factors 2, 3, and 4, respectively, on datasets Set5, Set14, B100, Urban100, and Manga109 of the proposed method and other lightweight models.…”

Section: Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

“…Then, in the IMDN, 14 information multi-distillate blocks were designed to segment and progressively refine features to reduce the amount of computation and aggregate them according to the importance of candidate features. Furthermore, to better integrate local-global information, Zeng et al 19 proposed the lightweight global-locally connected distillation network (GLCDNet). The pixel attention networks 20 proposed by Zhao et al introduced the pixel attention mechanism and self-calibrating convolution to achieve better reconstruction performance with smaller parameters.…”

Section: Lightweight Sr Reconstruction Modelmentioning

confidence: 99%

See 2 more Smart Citations

Lightweight image super-resolution reconstruction based on inverted residual attention network

et al. 2023

View full text Add to dashboard Cite

.In recent years, methods based on deep convolutional neural networks have made great progress in the field of image super-resolution reconstruction. However, mainstream approaches generally establish many network layers, leading to high computational costs and memory usage that are unsuitable for resource-limited edge devices. To alleviate this issue, a lightweight inverted residual attention network (IRAN) is proposed to obtain better super-resolution reconstruction performance with lower parameters and computation. The main structure of the IRAN consists of a series of inverted residual attention groups (IRAGs), which are mainly composed of several inverted residual attention blocks (IRABs). IRAB effectively reduces the network parameters and computation while extracting depth features by introducing the inverted residual structure and uses the simple channel attention mechanism to learn the important channel feature information. In addition, an enhanced spatial attention mechanism is introduced at the beginning and end of IRAG to further improve the reconstruction performance of the network. The experimental results show that compared with the mainstream lightweight networks, not only the peak signal-to-noise ratio and structural similarity of image quality metrics are better but also the parameters and the computational effort of the network are lower.

show abstract