Salient object detection via reciprocal function filter

Wenbing Chen, Keshav Dahal, Shuxian Huang

Research output: Contribution to journalArticle

105 Downloads (Pure)

Abstract

Salient object detection (SOD) has been attracting a lot of interest, and recently many computational models have been developed. In this paper, we formulate a SOD model, in which saliency map is computed as a combination of the colour, its distribution-based saliency and orientation saliency. Similar to traditional SODs, the proposed method is based on super-pixel segmentation and super-pixelutilizes both colour and its distribution-based saliency to generate a coarse saliency map. However, distinct from traditional SODs, we further use orientation contrast to optimize the coarse saliency map to obtain an improved saliency map. Our contributions are twofold. First, we combine colour uniqueness and its distribution with local orientation information (LOI) used in Itti’s model to effectively improve profiles of salient regions. Second, a reciprocal function is defined to substitute the Gabor function used in LOI and we have proved that the substitution could detect relatively homogeneous and uniform regions at the boundary of salient object, whereas it is what the traditional models lack. Our approach significantly outperforms state-of-the-art methods on four benchmark datasets while, we demonstrate that the proposed method runs as fast as most existing algorithms.
Original languageEnglish
Pages (from-to)1616-1624
Number of pages9
JournalIET Image Processing
Volume13
Issue number10
Early online date18 Apr 2019
DOIs
Publication statusE-pub ahead of print - 18 Apr 2019

Fingerprint

Color
Substitution reactions
Pixels
Object detection

Cite this

Chen, Wenbing ; Dahal, Keshav ; Huang, Shuxian. / Salient object detection via reciprocal function filter. In: IET Image Processing. 2019 ; Vol. 13, No. 10. pp. 1616-1624.
@article{6c2785edcc2043d4b55e2c2a2dff4944,
title = "Salient object detection via reciprocal function filter",
abstract = "Salient object detection (SOD) has been attracting a lot of interest, and recently many computational models have been developed. In this paper, we formulate a SOD model, in which saliency map is computed as a combination of the colour, its distribution-based saliency and orientation saliency. Similar to traditional SODs, the proposed method is based on super-pixel segmentation and super-pixelutilizes both colour and its distribution-based saliency to generate a coarse saliency map. However, distinct from traditional SODs, we further use orientation contrast to optimize the coarse saliency map to obtain an improved saliency map. Our contributions are twofold. First, we combine colour uniqueness and its distribution with local orientation information (LOI) used in Itti’s model to effectively improve profiles of salient regions. Second, a reciprocal function is defined to substitute the Gabor function used in LOI and we have proved that the substitution could detect relatively homogeneous and uniform regions at the boundary of salient object, whereas it is what the traditional models lack. Our approach significantly outperforms state-of-the-art methods on four benchmark datasets while, we demonstrate that the proposed method runs as fast as most existing algorithms.",
author = "Wenbing Chen and Keshav Dahal and Shuxian Huang",
year = "2019",
month = "4",
day = "18",
doi = "10.1049/iet-ipr.2018.5722",
language = "English",
volume = "13",
pages = "1616--1624",
journal = "IET Image Processing",
issn = "1751-9659",
publisher = "Institution of Engineering and Technology",
number = "10",

}

Salient object detection via reciprocal function filter. / Chen, Wenbing; Dahal, Keshav; Huang, Shuxian.

In: IET Image Processing, Vol. 13, No. 10, 18.04.2019, p. 1616-1624.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Salient object detection via reciprocal function filter

AU - Chen, Wenbing

AU - Dahal, Keshav

AU - Huang, Shuxian

PY - 2019/4/18

Y1 - 2019/4/18

N2 - Salient object detection (SOD) has been attracting a lot of interest, and recently many computational models have been developed. In this paper, we formulate a SOD model, in which saliency map is computed as a combination of the colour, its distribution-based saliency and orientation saliency. Similar to traditional SODs, the proposed method is based on super-pixel segmentation and super-pixelutilizes both colour and its distribution-based saliency to generate a coarse saliency map. However, distinct from traditional SODs, we further use orientation contrast to optimize the coarse saliency map to obtain an improved saliency map. Our contributions are twofold. First, we combine colour uniqueness and its distribution with local orientation information (LOI) used in Itti’s model to effectively improve profiles of salient regions. Second, a reciprocal function is defined to substitute the Gabor function used in LOI and we have proved that the substitution could detect relatively homogeneous and uniform regions at the boundary of salient object, whereas it is what the traditional models lack. Our approach significantly outperforms state-of-the-art methods on four benchmark datasets while, we demonstrate that the proposed method runs as fast as most existing algorithms.

AB - Salient object detection (SOD) has been attracting a lot of interest, and recently many computational models have been developed. In this paper, we formulate a SOD model, in which saliency map is computed as a combination of the colour, its distribution-based saliency and orientation saliency. Similar to traditional SODs, the proposed method is based on super-pixel segmentation and super-pixelutilizes both colour and its distribution-based saliency to generate a coarse saliency map. However, distinct from traditional SODs, we further use orientation contrast to optimize the coarse saliency map to obtain an improved saliency map. Our contributions are twofold. First, we combine colour uniqueness and its distribution with local orientation information (LOI) used in Itti’s model to effectively improve profiles of salient regions. Second, a reciprocal function is defined to substitute the Gabor function used in LOI and we have proved that the substitution could detect relatively homogeneous and uniform regions at the boundary of salient object, whereas it is what the traditional models lack. Our approach significantly outperforms state-of-the-art methods on four benchmark datasets while, we demonstrate that the proposed method runs as fast as most existing algorithms.

U2 - 10.1049/iet-ipr.2018.5722

DO - 10.1049/iet-ipr.2018.5722

M3 - Article

VL - 13

SP - 1616

EP - 1624

JO - IET Image Processing

JF - IET Image Processing

SN - 1751-9659

IS - 10

ER -