Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Memory-efficient divide-and-conquer attention for lightweight image super-resolution.

Rui He1, Zhenyang Zhu2, Xiaoyang Mao2

  • 1Integrated Graduate School of Medicine, Engineering, and Agricultural Sciences, University of Yamanashi, Address, kofu, Yamanashi, 400-8510, Japan.

Neural Networks : the Official Journal of the International Neural Network Society
|May 21, 2026
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Super-resolution Fluorescence Microscopy01:37

Super-resolution Fluorescence Microscopy

Super-resolution fluorescence microscopy (SRFM) provides a better resolution than conventional fluorescence microscopy by reducing the point spread function (PSF). PSF is the light intensity distribution from a point that causes it to appear blurred. Due to PSF, each fluorescing point appears bigger than its actual size, and it is the PSF interference of nearby fluorophores that causes the blurred image. Various approaches to achieving higher resolution through SRFM have recently been developed.

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Cardiovascular Complications in Hematopoietic Stem Cell Transplanted Patients.

Journal of personalized medicine·2022
Same author

Effect of Alkali and Sulfate on the Hydration Characteristic of Cement-Based Materials Containing Coal Gasification Slag.

Materials (Basel, Switzerland)·2022
Same author

Effects of Simultaneous Application of Double Chelating Agents to Pb-Contaminated Soil on the Phytoremediation Efficiency of <i>Indocalamus decorus</i> Q. H. Dai and the Soil Environment.

Toxics·2022
Same author

[Protective effect of active vitamin D on liver fibrosis induced by sodium arsenite in SD rats].

Wei sheng yan jiu = Journal of hygiene research·2022
Same author

Coronatine promotes maize water uptake by directly binding to the aquaporin ZmPIP2;5 and enhancing its activity.

Journal of integrative plant biology·2022
Same author

Hyperphosphorylation of EGFR/ERK signaling facilitates long-term arsenite-induced hepatocytes epithelial-mesenchymal transition and liver fibrosis in sprague-dawley rats.

Ecotoxicology and environmental safety·2022
Same journal

Q-learning based asynchronous Boolean control networks stabilization with data loss.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

New results on prescribed-time synchronization of complex networks via intermittent control.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Variance-constrained multi-view ensemble broad network for imbalanced data.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Dynamic analysis and reliable mechanical optimization application of ring HNN effected with a memristive neuron.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

DAFF-Net: A detection and search method for small-scale low surface brightness galaxies.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Quasi-synchronization for complex networks with hybrid pinning intermittent control.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

This study introduces Memory-Efficient Divide-and-Conquer Attention (MEDCA) for image super-resolution (SR), significantly cutting memory use. MEDCA achieves better reconstruction quality and speed, making SR practical for devices with limited resources.

Area of Science:

  • Computer Vision
  • Artificial Intelligence
  • Deep Learning

Background:

  • Transformer-based methods excel in lightweight image super-resolution (SR).
  • Existing SR methods often neglect memory consumption, limiting use on resource-constrained devices.
  • High memory and space complexity of standard window-based self-attention (WSA) hinder practical applications.

Purpose of the Study:

  • To propose a novel memory-efficient attention mechanism for SR.
  • To reduce memory usage without compromising reconstruction quality or inference speed.
  • To enhance the practicality of SR on devices with limited computational resources.

Main Methods:

  • Introduced Memory-Efficient Divide-and-Conquer Attention (MEDCA) for SR.
  • Employed a divide-and-conquer strategy by splitting input features into channel-wise subspaces.
Keywords:
Divide-and-conquerDividing strategiesLightweight image super-resolutionMemory-efficient

Related Experiment Videos

  • Applied asymmetric partitioning and self-attention within subspaces, followed by output aggregation.

    • Main Results:

    • MEDCA reduces space complexity by 50% compared to traditional WSA methods like SwinIR.
    • Achieved significant memory reduction (70.3% less than HiT-SRF) while improving performance by 0.12dB.
    • Maintained competitive inference speed across multiple benchmark datasets.

    Conclusions:

    • MEDCA offers substantial memory savings for SR tasks.
    • The proposed method enhances SR performance and maintains efficient inference.
    • MEDCA presents a practical solution for deploying SR on memory-limited devices.