On fast sample preselection for speeding up convolutional neural network training

Frédéric Rayar, Seiichi Uchida

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

Abstract

We propose a fast hybrid statistical and graph-based sample preselection method for speeding up CNN training process. To do so, we process each class separately: some candidates are first extracted based on their distances to the class mean. Then, we structure all the candidates in a graph representation and use it to extract the final set of preselected samples. The proposed method is evaluated and discussed based on an image classification task, on three data sets that contain up to several hundred thousands of images.

Original languageEnglish
Title of host publicationStructural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings
EditorsEdwin R. Hancock, Tin Kam Ho, Battista Biggio, Richard C. Wilson, Antonio Robles-Kelly, Xiao Bai
PublisherSpringer Verlag
Pages65-75
Number of pages11
ISBN (Print)9783319977843
DOIs
Publication statusPublished - Jan 1 2018
EventJoint IAPR International Workshops on Structural and Syntactic Pattern Recognition, SSPR 2018 and Statistical Techniques in Pattern Recognition, SPR 2018 - Beijing, China
Duration: Aug 17 2018Aug 19 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11004 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

OtherJoint IAPR International Workshops on Structural and Syntactic Pattern Recognition, SSPR 2018 and Statistical Techniques in Pattern Recognition, SPR 2018
CountryChina
CityBeijing
Period8/17/188/19/18

Fingerprint

Image classification
Neural Networks
Neural networks
Graph Representation
Image Classification
Graph in graph theory
Class
Training

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Rayar, F., & Uchida, S. (2018). On fast sample preselection for speeding up convolutional neural network training. In E. R. Hancock, T. K. Ho, B. Biggio, R. C. Wilson, A. Robles-Kelly, & X. Bai (Eds.), Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings (pp. 65-75). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11004 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-97785-0_7

On fast sample preselection for speeding up convolutional neural network training. / Rayar, Frédéric; Uchida, Seiichi.

Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings. ed. / Edwin R. Hancock; Tin Kam Ho; Battista Biggio; Richard C. Wilson; Antonio Robles-Kelly; Xiao Bai. Springer Verlag, 2018. p. 65-75 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11004 LNCS).

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

Rayar, F & Uchida, S 2018, On fast sample preselection for speeding up convolutional neural network training. in ER Hancock, TK Ho, B Biggio, RC Wilson, A Robles-Kelly & X Bai (eds), Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11004 LNCS, Springer Verlag, pp. 65-75, Joint IAPR International Workshops on Structural and Syntactic Pattern Recognition, SSPR 2018 and Statistical Techniques in Pattern Recognition, SPR 2018, Beijing, China, 8/17/18. https://doi.org/10.1007/978-3-319-97785-0_7
Rayar F, Uchida S. On fast sample preselection for speeding up convolutional neural network training. In Hancock ER, Ho TK, Biggio B, Wilson RC, Robles-Kelly A, Bai X, editors, Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings. Springer Verlag. 2018. p. 65-75. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-97785-0_7
Rayar, Frédéric ; Uchida, Seiichi. / On fast sample preselection for speeding up convolutional neural network training. Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings. editor / Edwin R. Hancock ; Tin Kam Ho ; Battista Biggio ; Richard C. Wilson ; Antonio Robles-Kelly ; Xiao Bai. Springer Verlag, 2018. pp. 65-75 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{9d72850368b14f4fa4a9b6a2f1eb786d,
title = "On fast sample preselection for speeding up convolutional neural network training",
abstract = "We propose a fast hybrid statistical and graph-based sample preselection method for speeding up CNN training process. To do so, we process each class separately: some candidates are first extracted based on their distances to the class mean. Then, we structure all the candidates in a graph representation and use it to extract the final set of preselected samples. The proposed method is evaluated and discussed based on an image classification task, on three data sets that contain up to several hundred thousands of images.",
author = "Fr{\'e}d{\'e}ric Rayar and Seiichi Uchida",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-97785-0_7",
language = "English",
isbn = "9783319977843",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "65--75",
editor = "Hancock, {Edwin R.} and Ho, {Tin Kam} and Battista Biggio and Wilson, {Richard C.} and Antonio Robles-Kelly and Xiao Bai",
booktitle = "Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings",
address = "Germany",

}

TY - GEN

T1 - On fast sample preselection for speeding up convolutional neural network training

AU - Rayar, Frédéric

AU - Uchida, Seiichi

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We propose a fast hybrid statistical and graph-based sample preselection method for speeding up CNN training process. To do so, we process each class separately: some candidates are first extracted based on their distances to the class mean. Then, we structure all the candidates in a graph representation and use it to extract the final set of preselected samples. The proposed method is evaluated and discussed based on an image classification task, on three data sets that contain up to several hundred thousands of images.

AB - We propose a fast hybrid statistical and graph-based sample preselection method for speeding up CNN training process. To do so, we process each class separately: some candidates are first extracted based on their distances to the class mean. Then, we structure all the candidates in a graph representation and use it to extract the final set of preselected samples. The proposed method is evaluated and discussed based on an image classification task, on three data sets that contain up to several hundred thousands of images.

UR - http://www.scopus.com/inward/record.url?scp=85052208916&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85052208916&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-97785-0_7

DO - 10.1007/978-3-319-97785-0_7

M3 - Conference contribution

AN - SCOPUS:85052208916

SN - 9783319977843

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 65

EP - 75

BT - Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings

A2 - Hancock, Edwin R.

A2 - Ho, Tin Kam

A2 - Biggio, Battista

A2 - Wilson, Richard C.

A2 - Robles-Kelly, Antonio

A2 - Bai, Xiao

PB - Springer Verlag

ER -