SalientShape: Group saliency in image collections

Ming Ming Cheng, Niloy J. Mitra, Sharon Xiaolei Huang, Shi Min Hu

Research output: Contribution to journalArticle

146 Citations (Scopus)

Abstract

Efficiently identifying salient objects in large image collections is essential for many applications including image retrieval, surveillance, image annotation, and object recognition. We propose a simple, fast, and effective algorithm for locating and segmenting salient objects by analysing image collections. As a key novelty, we introduce group saliency to achieve superior unsupervised salient object segmentation by extracting salient objects (in collections of pre-filtered images) that maximize between-image similarities and within-image distinctness. To evaluate our method, we construct a large benchmark dataset consisting of 15 K images across multiple categories with 6000+ pixel-accurate ground truth annotations for salient object regions where applicable. In all our tests, group saliency consistently outperforms state-of-the-art single-image saliency algorithms, resulting in both higher precision and better recall. Our algorithm successfully handles image collections, of an order larger than any existing benchmark datasets, consisting of diverse and heterogeneous images from various internet sources.

Original languageEnglish (US)
Pages (from-to)443-453
Number of pages11
JournalVisual Computer
Volume30
Issue number4
DOIs
StatePublished - Jan 1 2014

Fingerprint

Image recognition
Object recognition
Image retrieval
Pixels
Internet

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Vision and Pattern Recognition
  • Computer Graphics and Computer-Aided Design

Cite this

Cheng, Ming Ming ; Mitra, Niloy J. ; Huang, Sharon Xiaolei ; Hu, Shi Min. / SalientShape : Group saliency in image collections. In: Visual Computer. 2014 ; Vol. 30, No. 4. pp. 443-453.
@article{212fc31581174866898d77d7b1f219fc,
title = "SalientShape: Group saliency in image collections",
abstract = "Efficiently identifying salient objects in large image collections is essential for many applications including image retrieval, surveillance, image annotation, and object recognition. We propose a simple, fast, and effective algorithm for locating and segmenting salient objects by analysing image collections. As a key novelty, we introduce group saliency to achieve superior unsupervised salient object segmentation by extracting salient objects (in collections of pre-filtered images) that maximize between-image similarities and within-image distinctness. To evaluate our method, we construct a large benchmark dataset consisting of 15 K images across multiple categories with 6000+ pixel-accurate ground truth annotations for salient object regions where applicable. In all our tests, group saliency consistently outperforms state-of-the-art single-image saliency algorithms, resulting in both higher precision and better recall. Our algorithm successfully handles image collections, of an order larger than any existing benchmark datasets, consisting of diverse and heterogeneous images from various internet sources.",
author = "Cheng, {Ming Ming} and Mitra, {Niloy J.} and Huang, {Sharon Xiaolei} and Hu, {Shi Min}",
year = "2014",
month = "1",
day = "1",
doi = "10.1007/s00371-013-0867-4",
language = "English (US)",
volume = "30",
pages = "443--453",
journal = "Visual Computer",
issn = "0178-2789",
publisher = "Springer Verlag",
number = "4",

}

SalientShape : Group saliency in image collections. / Cheng, Ming Ming; Mitra, Niloy J.; Huang, Sharon Xiaolei; Hu, Shi Min.

In: Visual Computer, Vol. 30, No. 4, 01.01.2014, p. 443-453.

Research output: Contribution to journalArticle

TY - JOUR

T1 - SalientShape

T2 - Group saliency in image collections

AU - Cheng, Ming Ming

AU - Mitra, Niloy J.

AU - Huang, Sharon Xiaolei

AU - Hu, Shi Min

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Efficiently identifying salient objects in large image collections is essential for many applications including image retrieval, surveillance, image annotation, and object recognition. We propose a simple, fast, and effective algorithm for locating and segmenting salient objects by analysing image collections. As a key novelty, we introduce group saliency to achieve superior unsupervised salient object segmentation by extracting salient objects (in collections of pre-filtered images) that maximize between-image similarities and within-image distinctness. To evaluate our method, we construct a large benchmark dataset consisting of 15 K images across multiple categories with 6000+ pixel-accurate ground truth annotations for salient object regions where applicable. In all our tests, group saliency consistently outperforms state-of-the-art single-image saliency algorithms, resulting in both higher precision and better recall. Our algorithm successfully handles image collections, of an order larger than any existing benchmark datasets, consisting of diverse and heterogeneous images from various internet sources.

AB - Efficiently identifying salient objects in large image collections is essential for many applications including image retrieval, surveillance, image annotation, and object recognition. We propose a simple, fast, and effective algorithm for locating and segmenting salient objects by analysing image collections. As a key novelty, we introduce group saliency to achieve superior unsupervised salient object segmentation by extracting salient objects (in collections of pre-filtered images) that maximize between-image similarities and within-image distinctness. To evaluate our method, we construct a large benchmark dataset consisting of 15 K images across multiple categories with 6000+ pixel-accurate ground truth annotations for salient object regions where applicable. In all our tests, group saliency consistently outperforms state-of-the-art single-image saliency algorithms, resulting in both higher precision and better recall. Our algorithm successfully handles image collections, of an order larger than any existing benchmark datasets, consisting of diverse and heterogeneous images from various internet sources.

UR - http://www.scopus.com/inward/record.url?scp=84898597694&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84898597694&partnerID=8YFLogxK

U2 - 10.1007/s00371-013-0867-4

DO - 10.1007/s00371-013-0867-4

M3 - Article

AN - SCOPUS:84898597694

VL - 30

SP - 443

EP - 453

JO - Visual Computer

JF - Visual Computer

SN - 0178-2789

IS - 4

ER -