Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jul 13, 2025

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

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

Published on: December 15, 2023

SPA²Net: Structure-Preserved Attention Activated Network for Weakly Supervised Object Localization.

Dong Chen, Xingjia Pan, Fan Tang

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society

|October 17, 2023

Summary

This summary is machine-generated.

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

Cpeb4 regulates cardiomyocyte apoptosis in heart failure with association to Eif4a2 splicing modulation.

Scientific reports·2026

Same author

The N‑Glycoproteomic Landscape of the Lung in Monocrotaline-Induced Pulmonary Arterial Hypertension.

ACS omega·2026

Same author

ACE2 ameliorates DOX-induced cardiotoxicity by suppressing excessive autophagy via the AMPK/mTOR signaling pathway.

Biochemical pharmacology·2026

Same author

Analysis of the epidemiological features and factors associated with falls among the elderly in urban and rural areas of Chongqing, China: a cross-sectional study.

BMC public health·2026

Same author

Global, regional, and national trends in blindness and vision loss, 1990-2021: a secondary ecological trend analysis based on modelled population estimates.

Journal of global health·2026

Same author

SignMoD: Sign Language Video Generation via Mixture of Diffusion.

IEEE transactions on pattern analysis and machine intelligence·2026

Same journal

Style-Aware Contrastive Test-Time Adaptation: A Dual-Cache Model for Robust Vision-Language Alignment.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Semantic Frame Interpolation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Physics-Guided Cross-Modal Decoupling with Test-Time Adaptation for Hyperspectral Image Restoration.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

Change-Prior-Guided Unsupervised Change Detection of Heterogeneous Remote Sensing Images.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

AgonicDreamer: Enhancing Multi-View Consistency in Text-to-3D Generation via Rectified Score Distillation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same journal

BiCM-Prompt: Bidirectional Cross-Modal Prompt Tuning for Class-Incremental Learning on Multisource Remote Sensing Images.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

See all related articles

Structure-Preserved Attention Activated Network (SPA2Net) improves weakly supervised object localization (WSOL) by preserving structural information in deep features. This novel approach enhances accuracy by reducing partial activation issues common in classification-trained networks.

Area of Science:

Computer Vision
Deep Learning
Artificial Intelligence

Background:

Weakly supervised object localization (WSOL) methods utilize classification tasks to determine object positions in images.
A key challenge in WSOL is the partial activation problem, which hinders accurate object localization due to discriminant functions in deep Convolutional Neural Networks (CNNs).

Purpose of the Study:

To introduce a novel framework, SPA2Net, that addresses the partial activation problem in WSOL.
To enhance the structural preservation capabilities of deep features for more precise object localization.
To decouple the localization task from the classification branch for reduced mutual interference.

Main Methods:

SPA2Net employs a one-stage WSOL framework incorporating a structure-preserved attention mechanism.

More Related Videos

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

Published on: January 18, 2020

Related Experiment Videos

Last Updated: Jul 13, 2025

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

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

Published on: December 15, 2023

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

Published on: January 18, 2020

It utilizes high-order self-correlation as a structural prior to improve spatial interaction perception in convolutional features.

A self-supervised localization branch refines localization masks, and Restricted Activation Loss (RAL) distinguishes foreground from background.

The framework combines structural priors with spatial attention to promote activation spread from object parts to the whole.

Main Results:

SPA2Net demonstrates substantial and consistent performance improvements over baseline WSOL approaches.
The method effectively mitigates the partial activation problem, leading to more accurate object localization.
Experiments on CUB-200-2011 and ILSVRC benchmarks validate the efficacy of SPA2Net.

Conclusions:

SPA2Net offers a simple yet effective solution for accurate WSOL by leveraging structural priors and attention mechanisms.
The proposed framework successfully enhances object localization accuracy by preserving deep feature structures.
The decoupled localization branch and self-supervised refinement contribute to class-irrelevant localization map prediction.