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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Wireless Sensing and Networking for the Internet of Things.

Sensors (Basel, Switzerland)·2023
Same author

A Novel Floating High-Voltage Level Shifter with Pre-Storage Technique.

Sensors (Basel, Switzerland)·2022
Same author

Determining respiratory rate from photoplethysmogram and electrocardiogram signals using respiratory quality indices and neural networks.

PloS one·2021
Same author

Outcomes from international field trials with Male Aedes Sound Traps: Frequency-dependent effectiveness in capturing target species in relation to bycatch abundance.

PLoS neglected tropical diseases·2021
Same author

Optimizing spatial healthcare assets with Internet of Things.

Health information science and systems·2018

Related Experiment Video

Updated: Jan 7, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

990

Multi-Level Attribute-Guided-Based Adaptive Multi-Dilated Convolutional Network for Image Aesthetic Assessment.

Sumei Li1, Mingxuan Xie1, Wei Xiang2,3

  • 1School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China.

Journal of Imaging
|December 24, 2025
PubMed
Summary

This study introduces a novel Multi-level Attribute-Guided Adaptive Multi-Dilated Convolutional Network (MAADN) for image aesthetic assessment. MAADN improves accuracy by considering hierarchical attribute interactions and mitigating preprocessing distortions.

Keywords:
aesthetic attributesattention mechanismdilated convolutionimage aesthetics assessment

More Related Videos

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

721
Transient Optical Clearing Using Absorbing Molecules for Ex Vivo and In Vivo Imaging
07:15

Transient Optical Clearing Using Absorbing Molecules for Ex Vivo and In Vivo Imaging

Published on: July 11, 2025

2.2K

Related Experiment Videos

Last Updated: Jan 7, 2026

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

990
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

721
Transient Optical Clearing Using Absorbing Molecules for Ex Vivo and In Vivo Imaging
07:15

Transient Optical Clearing Using Absorbing Molecules for Ex Vivo and In Vivo Imaging

Published on: July 11, 2025

2.2K

Area of Science:

  • Computer Vision
  • Artificial Intelligence
  • Image Processing

Background:

  • Image aesthetic assessment (IAA) is vital for research and applications.
  • Existing IAA methods neglect hierarchical attribute-feature interactions and suffer from aspect ratio distortion during preprocessing, losing aesthetic information.

Purpose of the Study:

  • To propose a novel network, MAADN, that addresses limitations in IAA.
  • To leverage multi-level attribute features for guiding aesthetic assessment.
  • To reduce the negative impact of image preprocessing via adaptive dilated convolution.

Main Methods:

  • Developed a dual-branch architecture: one for multi-level attribute feature extraction, another for attribute-guided aesthetic feature learning.
  • Introduced an Attention-based Attribute-Guided Aesthetic Module (AGAM) using visual attention.
  • Designed an Adaptive Multi-Dilate Rate Convolution Module (AMDM) for adaptive fusion of dilated convolution features.

Main Results:

  • The proposed MAADN model significantly outperforms current state-of-the-art IAA approaches.
  • Experimental results validate the effectiveness of the dual-branch architecture and adaptive modules.
  • Visual analysis confirms MAADN's precise localization of aesthetically critical regions.

Conclusions:

  • MAADN effectively addresses hierarchical interactions and preprocessing distortions in IAA.
  • The adaptive nature of the proposed modules enhances flexibility and performance.
  • MAADN offers a robust and accurate solution for image aesthetic assessment.