A fast learning algorithm for image segmentation with max-pooling convolutional networks

Jonathan Masci; Alessandro Giusti; Dan Ciresan; Gabriel Fricout; Jurgen Schmidhuber

doi:10.1109/ICIP.2013.6738559

A fast learning algorithm for image segmentation with max-pooling convolutional networks

Jonathan Masci, Alessandro Giusti, Dan Ciresan, Gabriel Fricout, Jurgen Schmidhuber

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

49 Scopus citations

Abstract

We present a fast algorithm for training MaxPooling Convolutional Networks to segment images. This type of network yields record-breaking performance in a variety of tasks, but is normally trained on a computationally expensive patch-by-patch basis. Our new method processes each training image in a single pass, which is vastly more efficient. We validate the approach in different scenarios and report a 1500-fold speed-up. In an application to automated steel defect detection and segmentation, we obtain excellent performance with short training times. © 2013 IEEE.

Original language	English (US)
Title of host publication	2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings
Publisher	IEEE Computer Society
Pages	2713-2717
Number of pages	5
ISBN (Print)	9781479923410
DOIs	https://doi.org/10.1109/ICIP.2013.6738559
State	Published - Jan 1 2013
Externally published	Yes

Access to Document

10.1109/ICIP.2013.6738559

Cite this

@inproceedings{e45c649d161c40159ea0a5ec0e57e719,

title = "A fast learning algorithm for image segmentation with max-pooling convolutional networks",

abstract = "We present a fast algorithm for training MaxPooling Convolutional Networks to segment images. This type of network yields record-breaking performance in a variety of tasks, but is normally trained on a computationally expensive patch-by-patch basis. Our new method processes each training image in a single pass, which is vastly more efficient. We validate the approach in different scenarios and report a 1500-fold speed-up. In an application to automated steel defect detection and segmentation, we obtain excellent performance with short training times. {\textcopyright} 2013 IEEE.",

author = "Jonathan Masci and Alessandro Giusti and Dan Ciresan and Gabriel Fricout and Jurgen Schmidhuber",

note = "Generated from Scopus record by KAUST IRTS on 2022-09-14",

year = "2013",

month = jan,

day = "1",

doi = "10.1109/ICIP.2013.6738559",

language = "English (US)",

isbn = "9781479923410",

pages = "2713--2717",

booktitle = "2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings",

publisher = "IEEE Computer Society",

address = "United States",

}

TY - GEN

T1 - A fast learning algorithm for image segmentation with max-pooling convolutional networks

AU - Masci, Jonathan

AU - Giusti, Alessandro

AU - Ciresan, Dan

AU - Fricout, Gabriel

AU - Schmidhuber, Jurgen

N1 - Generated from Scopus record by KAUST IRTS on 2022-09-14

PY - 2013/1/1

Y1 - 2013/1/1

N2 - We present a fast algorithm for training MaxPooling Convolutional Networks to segment images. This type of network yields record-breaking performance in a variety of tasks, but is normally trained on a computationally expensive patch-by-patch basis. Our new method processes each training image in a single pass, which is vastly more efficient. We validate the approach in different scenarios and report a 1500-fold speed-up. In an application to automated steel defect detection and segmentation, we obtain excellent performance with short training times. © 2013 IEEE.

AB - We present a fast algorithm for training MaxPooling Convolutional Networks to segment images. This type of network yields record-breaking performance in a variety of tasks, but is normally trained on a computationally expensive patch-by-patch basis. Our new method processes each training image in a single pass, which is vastly more efficient. We validate the approach in different scenarios and report a 1500-fold speed-up. In an application to automated steel defect detection and segmentation, we obtain excellent performance with short training times. © 2013 IEEE.

UR - http://ieeexplore.ieee.org/document/6738559/

UR - http://www.scopus.com/inward/record.url?scp=84897756043&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2013.6738559

DO - 10.1109/ICIP.2013.6738559

M3 - Conference contribution

SN - 9781479923410

SP - 2713

EP - 2717

BT - 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings

PB - IEEE Computer Society

ER -

A fast learning algorithm for image segmentation with max-pooling convolutional networks

Abstract

Access to Document

Other files and links

Fingerprint

Cite this