EUVIMEDEUVIMED entdecken
Crossref

Wavelet-Aided Directional Convolutional Network for Image Super-Resolution

Ruilin Zhang, Yongsheng Dong, Yang Zhao, Songtao Liu

IEEE Signal Processing Letters · 2026 · 33 · 2630-2634

DOI: 10.1109/lsp.2026.3702797 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Ruilin Zhang und 3 weiteren Autor:innen. Die Arbeit nennt 29 Literaturangaben. Gefördert von: Natural Science Foundation of Henan Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Although small kernel convolution is widely used in single image super-resolution, it struggles with capturing high- and low-frequency features due to limited local aggregation and remote dependency modeling. Moreover, traditional spatial transformations lack learnable feature combination. To alleviate these issues, in this letter we propose a wavelet-aided directional convolutional network (WADCNet) for image super-resolution. Our proposed pixel directional convolution (PDC) module employs an adaptive directional mask convolution operation to process pixels and enhance local feature aggregation and information fusion. Our proposed retain high block (RHB) module is mainly based on image brightening and high-frequency retention (HFR), capture more edge and texture features and preserve high-frequency information in the image. Finally, the wavelet transform is combined with the network as the parallel upsampling path to generate the final super-resolution output image. Experimental results demonstrate that our proposed WADCNet is effective when compared with the existing low-parameter super-resolution methods.

Förderung

  • Natural Science Foundation of Henan Province

Zitieren

APA

Zhang, R., Dong, Y., Zhao, Y., & Liu, S. (2026). Wavelet-Aided Directional Convolutional Network for Image Super-Resolution. IEEE Signal Processing Letters, 33, 2630-2634. https://doi.org/10.1109/lsp.2026.3702797

Vancouver

Zhang R, Dong Y, Zhao Y, Liu S. Wavelet-Aided Directional Convolutional Network for Image Super-Resolution. IEEE Signal Processing Letters. 2026;33:2630-2634. doi:10.1109/lsp.2026.3702797

BibTeX
@article{Zhang2026,
  author = {Zhang, Ruilin and Dong, Yongsheng and Zhao, Yang and Liu, Songtao},
  title = {Wavelet-Aided Directional Convolutional Network for Image Super-Resolution},
  journal = {IEEE Signal Processing Letters},
  year = {2026},
  volume = {33},
  pages = {2630--2634},
  doi = {10.1109/lsp.2026.3702797}
}
Kostenlos starten →