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: Mar 16, 2026

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

Rank-based pooling for deep convolutional neural networks.

Zenglin Shi¹, Yangdong Ye¹, Yunpeng Wu¹

¹School of Information Engineering, Zhengzhou University, Zhengzhou, 450052, China.

Neural Networks : the Official Journal of the International Neural Network Society

|August 21, 2016

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

Association of age and primary treatment with risk of non-lymphoma-related death and long-term survival outcomes in adult patients with early-stage follicular lymphoma: a population-based analysis.

Annals of hematology·2026

Same author

Learning Dual Transformers for All-in-One Image Restoration From a Frequency Perspective.

IEEE transactions on neural networks and learning systems·2026

Same author

Inhibitory effects of esculetin as a quorum sensing inhibitor on biofilm formation and virulence factors in Vibrio anguillarum.

World journal of microbiology & biotechnology·2026

Same author

Discovery of novel CSF1R inhibitor for triple-negative breast cancer (TNBC) treatment through TAMs reprogramming.

Biochemical pharmacology·2026

Same author

Evolutionary insights and structural characterization guide the development of RAG1/RAG2-deficient swine models for immunological research.

Frontiers in immunology·2026

Same author

Compact and informative representation learning for scRNA-seq data clustering with masked information bottleneck.

BMC biology·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

Same journal

Physics-encoded convolutional neural operators for parametric PDEs: A convergence-guaranteed framework via pre-computed kernel fields.

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

Same journal

Exploiting audio-visual modalities in videos: Object detection via multi-stage bilateral coupling network.

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

Same journal

Reliability-aware modality completion with cross-modal distillation for federated learning with missing modalities.

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

See all related articles

Researchers introduce rank-based pooling for deep convolutional neural networks (CNNs), offering a robust alternative to traditional methods. This novel approach enhances classification performance and addresses scale issues in image recognition tasks.

Area of Science:

Computer Vision
Machine Learning
Deep Learning

Background:

Pooling is crucial for translation invariance in Convolutional Neural Networks (CNNs).
Conventional pooling methods operate on activation values, which can be sensitive to scale variations.
Existing pooling techniques consistently improve CNN performance but have limitations.

Purpose of the Study:

To propose a novel rank-based pooling mechanism for CNNs.
To enhance the robustness and performance of pooling operations.
To introduce a new criterion for analyzing the discriminant ability of pooling methods.

Main Methods:

Developed rank-based pooling, leveraging the invariance of ranking lists to activation value changes.
Introduced rank-based average pooling (RAP), rank-based weighted pooling (RWP), and rank-based stochastic pooling (RSP).

Keywords:

Convolutional neural network Deep learning Image classification Pooling

More Related Videos

Deep Neural Networks for Image-Based Dietary Assessment

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

Related Experiment Videos

Last Updated: Mar 16, 2026

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

Deep Neural Networks for Image-Based Dietary Assessment

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

Proposed a novel criterion to analyze the discriminant ability of pooling methods.

Main Results:

Rank-based pooling demonstrates superior classification performance compared to existing methods on image benchmarks.
Rank-based stochastic pooling (RSP) integration with Network-in-Network improved performance on CIFAR datasets.
The proposed rank-based approach offers improved robustness and avoids scale problems inherent in value-based methods.

Conclusions:

Rank-based pooling is a promising alternative to conventional pooling in CNNs.
The novel pooling mechanism enhances robustness and classification accuracy.
Further research into rank-based pooling and discriminant ability analysis is warranted.