A combined convolutional neural network and potential region-of-interest model for saliency detection

Yu HU; Zhen LIANG; Zheru CHI; Hong FU

doi:10.1109/ICNC.2015.7377982

A combined convolutional neural network and potential region-of-interest model for saliency detection

Yu HU
, Zhen LIANG
, Zheru CHI
, Hong FU

Department of Mathematics and Information Technology (MIT)

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

1 Citation (Scopus)

Abstract

A saliency detection model for approaching the human performance is a challenging research topic. In this paper, a new saliency model is proposed to detect saliency in natural scenes by using a trained convolutional neural network and a region-based validation method. The convolutional neural network (CNN) focuses on image details and local contrast of an image, while the region-based validation method focus on global information. Experimental results show that the two components of the model are complementary for each other in producing high-quality saliency maps. Copyright © 2015 IEEE.

Original language	English
Title of host publication	2015 11th International Conference on Natural Computation (ICNC 2015)
Editors	Zheng XIAO, Zhao TONG, Kenli II, Xingwei WANG, Keqin LI
Place of Publication	Piscataway, NJ
Publisher	IEEE
Pages	154-158
ISBN (Electronic)	9781467376785, 9781467376792
DOIs	https://doi.org/10.1109/ICNC.2015.7377982
Publication status	Published - 2016

Keywords

Saliency map
Saliency detection
Machine learning
Convolutional neural networks

Access to Document

10.1109/ICNC.2015.7377982

A combined convolutional neural network and potential region-of-interest model for saliency detection. / HU, Yu; LIANG, Zhen; CHI, Zheru et al.
2015 11th International Conference on Natural Computation (ICNC 2015). ed. / Zheng XIAO; Zhao TONG; Kenli II; Xingwei WANG; Keqin LI. Piscataway, NJ: IEEE, 2016. p. 154-158.

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

@inbook{8efb0f8ce7944a2abea7b39a5771b1ad,

title = "A combined convolutional neural network and potential region-of-interest model for saliency detection",

abstract = "A saliency detection model for approaching the human performance is a challenging research topic. In this paper, a new saliency model is proposed to detect saliency in natural scenes by using a trained convolutional neural network and a region-based validation method. The convolutional neural network (CNN) focuses on image details and local contrast of an image, while the region-based validation method focus on global information. Experimental results show that the two components of the model are complementary for each other in producing high-quality saliency maps. Copyright {\textcopyright} 2015 IEEE.",

author = "Yu HU and Zhen LIANG and Zheru CHI and Hong FU",

year = "2016",

doi = "10.1109/ICNC.2015.7377982",

language = "English",

pages = "154--158",

editor = "Zheng XIAO and Zhao TONG and Kenli II and Xingwei WANG and Keqin LI",

booktitle = "2015 11th International Conference on Natural Computation (ICNC 2015)",

publisher = "IEEE",

}

TY - CHAP

T1 - A combined convolutional neural network and potential region-of-interest model for saliency detection

AU - HU, Yu

AU - LIANG, Zhen

AU - CHI, Zheru

AU - FU, Hong

PY - 2016

Y1 - 2016

N2 - A saliency detection model for approaching the human performance is a challenging research topic. In this paper, a new saliency model is proposed to detect saliency in natural scenes by using a trained convolutional neural network and a region-based validation method. The convolutional neural network (CNN) focuses on image details and local contrast of an image, while the region-based validation method focus on global information. Experimental results show that the two components of the model are complementary for each other in producing high-quality saliency maps. Copyright © 2015 IEEE.

AB - A saliency detection model for approaching the human performance is a challenging research topic. In this paper, a new saliency model is proposed to detect saliency in natural scenes by using a trained convolutional neural network and a region-based validation method. The convolutional neural network (CNN) focuses on image details and local contrast of an image, while the region-based validation method focus on global information. Experimental results show that the two components of the model are complementary for each other in producing high-quality saliency maps. Copyright © 2015 IEEE.

U2 - 10.1109/ICNC.2015.7377982

DO - 10.1109/ICNC.2015.7377982

M3 - Chapters

SP - 154

EP - 158

BT - 2015 11th International Conference on Natural Computation (ICNC 2015)

A2 - XIAO, Zheng

A2 - TONG, Zhao

A2 - II, Kenli

A2 - WANG, Xingwei

A2 - LI, Keqin

PB - IEEE

CY - Piscataway, NJ

ER -

A combined convolutional neural network and potential region-of-interest model for saliency detection

Abstract

Keywords

Access to Document

Fingerprint