On fast sample preselection for speeding up convolutional neural network training

Frédéric Rayar, Seiichi Uchida

研究成果: 著書/レポートタイプへの貢献会議での発言

抄録

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.

元の言語英語
ホスト出版物のタイトルStructural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, S+SSPR 2018, Proceedings
編集者Edwin R. Hancock, Tin Kam Ho, Battista Biggio, Richard C. Wilson, Antonio Robles-Kelly, Xiao Bai
出版者Springer Verlag
ページ65-75
ページ数11
ISBN(印刷物)9783319977843
DOI
出版物ステータス出版済み - 1 1 2018
イベントJoint IAPR International Workshops on Structural and Syntactic Pattern Recognition, SSPR 2018 and Statistical Techniques in Pattern Recognition, SPR 2018 - Beijing, 中国
継続期間: 8 17 20188 19 2018

出版物シリーズ

名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
11004 LNCS
ISSN(印刷物)0302-9743
ISSN(電子版)1611-3349

その他

その他Joint IAPR International Workshops on Structural and Syntactic Pattern Recognition, SSPR 2018 and Statistical Techniques in Pattern Recognition, SPR 2018
中国
Beijing
期間8/17/188/19/18

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)

これを引用

Rayar, F., & Uchida, S. (2018). On fast sample preselection for speeding up convolutional neural network training. : E. R. Hancock, T. K. Ho, B. Biggio, R. C. Wilson, A. Robles-Kelly, & X. Bai (版), 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); 巻数 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. 版 / 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); 巻 11004 LNCS).

研究成果: 著書/レポートタイプへの貢献会議での発言

Rayar, F & Uchida, S 2018, On fast sample preselection for speeding up convolutional neural network training. : ER Hancock, TK Ho, B Biggio, RC Wilson, A Robles-Kelly & X Bai (版), 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), 巻. 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, 中国, 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. : Hancock ER, Ho TK, Biggio B, Wilson RC, Robles-Kelly A, Bai X, 編集者, 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. 編集者 / 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 -