CNN training with graph-based sample preselection: application to handwritten character recognition

Frederic Rayar, Masanori Goto, Seiichi Uchida

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

In this paper, we present a study on sample preselection in large training data set for CNN-based classification. To do so, we structure the input data set in a network representation, namely the Relative Neighbourhood Graph, and then extract some vectors of interest. The proposed preselection method is evaluated in the context of handwritten character recognition, by using two data sets, up to several hundred thousands of images. It is shown that the graph-based preselection can reduce the training data set without degrading the recognition accuracy of a non pretrained CNN shallow model.

Original languageEnglish
Title of host publicationProceedings - 13th IAPR International Workshop on Document Analysis Systems, DAS 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages19-24
Number of pages6
ISBN (Electronic)9781538633465
DOIs
Publication statusPublished - Jun 22 2018
Event13th IAPR International Workshop on Document Analysis Systems, DAS 2018 - Vienna, Austria
Duration: Apr 24 2018Apr 27 2018

Publication series

NameProceedings - 13th IAPR International Workshop on Document Analysis Systems, DAS 2018

Other

Other13th IAPR International Workshop on Document Analysis Systems, DAS 2018
Country/TerritoryAustria
CityVienna
Period4/24/184/27/18

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition
  • Signal Processing

Fingerprint

Dive into the research topics of 'CNN training with graph-based sample preselection: application to handwritten character recognition'. Together they form a unique fingerprint.

Cite this