Object recognition with gradient-based learning

Yann LeCun, Patrick Haffner, Léon Bottou, Yoshua Bengio

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

Abstract

Finding an appropriate set of features s an essential problem in the design of shape recognition systems. This paper attempts to show that for recognizing simple objects with high shape variability such as handwritten characters, it is possible, and even advantageous, to feed the system directly with minimally processed images and to rely on learning to extract the right set of features. Convolutional Neural Networks are shown to be particularly well suited to this task. We also show that these networks can be used to recognize multiple objects without requiring explicit segmentation of the objects from their surrounding. The second part of the paper presents the Graph Transformer Network model which extends the applicability of gradient-based learning to systems that use graphs to represents features, objects, and their combinations.

Original languageEnglish (US)
Title of host publicationShape, Contour and Grouping in Computer Vision
EditorsDavid A. Forsyth, Joseph L. Mundy, Vito di Gesu, Roberto Cipolla
PublisherSpringer Verlag
Pages319-345
Number of pages27
ISBN (Print)3540667229, 9783540667223
DOIs
StatePublished - 1999
EventInternational Workshop on Shape, Contour and Grouping in Computer Vision - Palermo, Sicily, Italy
Duration: May 26 1998May 29 1998

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1681
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

OtherInternational Workshop on Shape, Contour and Grouping in Computer Vision
Country/TerritoryItaly
CityPalermo, Sicily
Period5/26/985/29/98

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Object recognition with gradient-based learning'. Together they form a unique fingerprint.

Cite this