Fully Mixed Max-Average Pooling: A Comparative Study for a Convolutional Neural Network

Brahim Ait Skourt, Nikola S. Nikolov, Aicha Majda

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

As an integral part of a convolutional neural network, the pooling layers are responsible for the down-sampling operation which aims at preventing overfitting. Besides the conventional pooling methods (max and average), various methods that involve mixing max pooling and average pooling have been proposed. In this work, we propose a new mixed pooling method, called fully mixed max-average pooling (FMMAP), and evaluate its performance within a comparative study of various conventional and state-of-the-art pooling methods. FMMAP consists of fully mixing max pooling and average pooling features instead of stochastically selecting them as done in other popular mixed-pooling methods. The experimental results suggest that FMMAP outperforms the conventional pooling methods in accuracy, and while being very close (within 0.05%) to the other mixed-pooling methods accuracy-wise, it significantly outperforms them in terms of running time, being at least 1.7 times and up to 2.7 times faster.

Original languageEnglish
Title of host publication2024 International Conference on Intelligent Systems and Computer Vision, ISCV 2024
EditorsMy Abdelouahed Sabri, Ali Yahyaouy, Khalid el Fazazy, Jamal Riffi, Mohamed Adnane Mahraz
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350350180
Publication statusPublished - 2024
Event6th International Conference on Intelligent Systems and Computer Vision, ISCV 2024 - Fez, Morocco
Duration: 8 May 202410 May 2024

Publication series

Name2024 International Conference on Intelligent Systems and Computer Vision, ISCV 2024

Conference

Conference6th International Conference on Intelligent Systems and Computer Vision, ISCV 2024
Country/TerritoryMorocco
CityFez
Period8/05/2410/05/24

Keywords

  • Average Pooling
  • Convolutional Neural Networks
  • Max Pooling
  • Mixed Pooling

Fingerprint

Dive into the research topics of 'Fully Mixed Max-Average Pooling: A Comparative Study for a Convolutional Neural Network'. Together they form a unique fingerprint.

Cite this