Feature pooling in scene character recognition: a comprehensive study

Zhong Zhang*, Hong Wang, Shuang Liu, Yunxue Shao

*Corresponding author for this work

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

Abstract

In this paper, we focus on the feature pooling methods for scene character recognition. We research three kinds of pooling methods: the average (sum) pooling, max pooling and weighted-based pooling methods. Specifically, various feature pooling methods are introduced, their merits and demerits are studied, and existing problems are discussed. Finally, we offer a specific comparison on the ICDAR2003 and Chars74k databases.

Original languageEnglish
Title of host publicationCommunications, Signal Processing, and Systems - Proceedings of the 2017 International Conference on Communications, Signal Processing, and Systems
EditorsQilian Liang, Min Jia, Jiasong Mu, Wei Wang, Xuhong Feng, Baoju Zhang
PublisherSpringer Verlag
Pages2150-2157
Number of pages8
ISBN (Print)9789811065705
DOIs
Publication statusPublished - 2019
Externally publishedYes
Event6th International Conference on Communications, Signal Processing, and Systems, CSPS 2017 - Harbin, China
Duration: 14 Jul 201716 Jul 2017

Publication series

NameLecture Notes in Electrical Engineering
Volume463
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

Conference6th International Conference on Communications, Signal Processing, and Systems, CSPS 2017
Country/TerritoryChina
CityHarbin
Period14/07/1716/07/17

Keywords

  • Feature pooling
  • Feature representation
  • Scene character recognition

Fingerprint

Dive into the research topics of 'Feature pooling in scene character recognition: a comprehensive study'. Together they form a unique fingerprint.

Cite this