On generalization of deep neural networks for visual recognition tasks

Takayuki Okatani

On generalization of deep neural networks for visual recognition tasks

INF - System Information Sciences

Research output: Contribution to journal › Conference article › peer-review

Abstract

This article discusses generalization ability of deep neural networks (DNNs) for visual recognition. It is known that DNNs easily fail for images to which noises are added, when they have not learned the noisy images. We discuss how to cope with such limitation of DNNs.

Original language	English
Pages (from-to)	961-964
Number of pages	4
Journal	Proceedings of the International Display Workshops
Volume	27
Publication status	Published - 2021 Dec 9
Event	27th International Display Workshops, IDW 2020 - Virtual, Online Duration: 2020 Dec 9 → 2020 Dec 11

Keywords

Computer vision
Deep learning
Image restoration
Out-of-distribution detection

ASJC Scopus subject areas

Computer Vision and Pattern Recognition
Human-Computer Interaction
Electrical and Electronic Engineering
Electronic, Optical and Magnetic Materials

Cite this

@article{56c80f1203074905b398c34e7e3d8b35,

title = "On generalization of deep neural networks for visual recognition tasks",

abstract = "This article discusses generalization ability of deep neural networks (DNNs) for visual recognition. It is known that DNNs easily fail for images to which noises are added, when they have not learned the noisy images. We discuss how to cope with such limitation of DNNs.",

keywords = "Computer vision, Deep learning, Image restoration, Out-of-distribution detection",

author = "Takayuki Okatani",

note = "Publisher Copyright: {\textcopyright} 2020 ITE and SID.; 27th International Display Workshops, IDW 2020 ; Conference date: 09-12-2020 Through 11-12-2020",

year = "2021",

month = dec,

day = "9",

language = "English",

volume = "27",

pages = "961--964",

journal = "Proceedings of the International Display Workshops",

issn = "1883-2490",

}

TY - JOUR

T1 - On generalization of deep neural networks for visual recognition tasks

AU - Okatani, Takayuki

PY - 2021/12/9

Y1 - 2021/12/9

N2 - This article discusses generalization ability of deep neural networks (DNNs) for visual recognition. It is known that DNNs easily fail for images to which noises are added, when they have not learned the noisy images. We discuss how to cope with such limitation of DNNs.

AB - This article discusses generalization ability of deep neural networks (DNNs) for visual recognition. It is known that DNNs easily fail for images to which noises are added, when they have not learned the noisy images. We discuss how to cope with such limitation of DNNs.

KW - Computer vision

KW - Deep learning

KW - Image restoration

KW - Out-of-distribution detection

UR - http://www.scopus.com/inward/record.url?scp=85119969894&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85119969894&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85119969894

VL - 27

SP - 961

EP - 964

JO - Proceedings of the International Display Workshops

JF - Proceedings of the International Display Workshops

SN - 1883-2490

T2 - 27th International Display Workshops, IDW 2020

Y2 - 9 December 2020 through 11 December 2020

ER -

On generalization of deep neural networks for visual recognition tasks

Abstract

Keywords

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this