Can the high-level content of natural images be indexed using local analysis?

John A. Black, Mariano Phielipp, Greg Nielson, Sethuraman Panchanathan

Research output: Contribution to journalArticle

Abstract

Early methods of image indexing relied heavily on color histograms, which characterize the global content of images. However, global indexing methods proved to be unsatisfactory, and researchers now employ more localized measures of image content, based on relatively small regions. At the same time, it has also become clear that image indexing should be based on higher-level visual content. This raises an important question: "Can the higher-level content of images be reliably indexed using local analysis?" In general, humans are better at indexing mid-level and high-level visual content than today's automated indexing algorithms. Therefore, it makes sense to ascertain how well humans can perform midlevel or high-level indexing, based on small regions. This paper describes research that employs a set of outdoor scenery images (called the NaturePix image set) to compare how successfully humans can label the visual content of small regions of natural images when (1) these regions are seen in the context of the larger image, and (2) when these regions are extracted from (and are seen in isolation from) that larger image. The results of these experiments indicate what types of higher-level image content can be recognized locally, and how successfully high-level image content can be indexed on the basis of local feature analysis.

Original languageEnglish (US)
Pages (from-to)414-425
Number of pages12
JournalUnknown Journal
Volume5292
DOIs
StatePublished - 2004

Fingerprint

Labels
Color
Experiments
Research Personnel
Research
histograms
isolation
color

Keywords

  • Content based image retrieval
  • Feature detectors
  • Image content
  • Image indexing
  • Lexical basis functions
  • Local content analysis
  • NaturePix
  • Semantic content
  • Semantic indexing
  • Visual content

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Can the high-level content of natural images be indexed using local analysis? / Black, John A.; Phielipp, Mariano; Nielson, Greg; Panchanathan, Sethuraman.

In: Unknown Journal, Vol. 5292, 2004, p. 414-425.

Research output: Contribution to journalArticle

Black, John A. ; Phielipp, Mariano ; Nielson, Greg ; Panchanathan, Sethuraman. / Can the high-level content of natural images be indexed using local analysis?. In: Unknown Journal. 2004 ; Vol. 5292. pp. 414-425.
@article{ed99427744b84d3e8d83af8577234de6,
title = "Can the high-level content of natural images be indexed using local analysis?",
abstract = "Early methods of image indexing relied heavily on color histograms, which characterize the global content of images. However, global indexing methods proved to be unsatisfactory, and researchers now employ more localized measures of image content, based on relatively small regions. At the same time, it has also become clear that image indexing should be based on higher-level visual content. This raises an important question: {"}Can the higher-level content of images be reliably indexed using local analysis?{"} In general, humans are better at indexing mid-level and high-level visual content than today's automated indexing algorithms. Therefore, it makes sense to ascertain how well humans can perform midlevel or high-level indexing, based on small regions. This paper describes research that employs a set of outdoor scenery images (called the NaturePix image set) to compare how successfully humans can label the visual content of small regions of natural images when (1) these regions are seen in the context of the larger image, and (2) when these regions are extracted from (and are seen in isolation from) that larger image. The results of these experiments indicate what types of higher-level image content can be recognized locally, and how successfully high-level image content can be indexed on the basis of local feature analysis.",
keywords = "Content based image retrieval, Feature detectors, Image content, Image indexing, Lexical basis functions, Local content analysis, NaturePix, Semantic content, Semantic indexing, Visual content",
author = "Black, {John A.} and Mariano Phielipp and Greg Nielson and Sethuraman Panchanathan",
year = "2004",
doi = "10.1117/12.527316",
language = "English (US)",
volume = "5292",
pages = "414--425",
journal = "Scanning Electron Microscopy",
issn = "0586-5581",
publisher = "Scanning Microscopy International",

}

TY - JOUR

T1 - Can the high-level content of natural images be indexed using local analysis?

AU - Black, John A.

AU - Phielipp, Mariano

AU - Nielson, Greg

AU - Panchanathan, Sethuraman

PY - 2004

Y1 - 2004

N2 - Early methods of image indexing relied heavily on color histograms, which characterize the global content of images. However, global indexing methods proved to be unsatisfactory, and researchers now employ more localized measures of image content, based on relatively small regions. At the same time, it has also become clear that image indexing should be based on higher-level visual content. This raises an important question: "Can the higher-level content of images be reliably indexed using local analysis?" In general, humans are better at indexing mid-level and high-level visual content than today's automated indexing algorithms. Therefore, it makes sense to ascertain how well humans can perform midlevel or high-level indexing, based on small regions. This paper describes research that employs a set of outdoor scenery images (called the NaturePix image set) to compare how successfully humans can label the visual content of small regions of natural images when (1) these regions are seen in the context of the larger image, and (2) when these regions are extracted from (and are seen in isolation from) that larger image. The results of these experiments indicate what types of higher-level image content can be recognized locally, and how successfully high-level image content can be indexed on the basis of local feature analysis.

AB - Early methods of image indexing relied heavily on color histograms, which characterize the global content of images. However, global indexing methods proved to be unsatisfactory, and researchers now employ more localized measures of image content, based on relatively small regions. At the same time, it has also become clear that image indexing should be based on higher-level visual content. This raises an important question: "Can the higher-level content of images be reliably indexed using local analysis?" In general, humans are better at indexing mid-level and high-level visual content than today's automated indexing algorithms. Therefore, it makes sense to ascertain how well humans can perform midlevel or high-level indexing, based on small regions. This paper describes research that employs a set of outdoor scenery images (called the NaturePix image set) to compare how successfully humans can label the visual content of small regions of natural images when (1) these regions are seen in the context of the larger image, and (2) when these regions are extracted from (and are seen in isolation from) that larger image. The results of these experiments indicate what types of higher-level image content can be recognized locally, and how successfully high-level image content can be indexed on the basis of local feature analysis.

KW - Content based image retrieval

KW - Feature detectors

KW - Image content

KW - Image indexing

KW - Lexical basis functions

KW - Local content analysis

KW - NaturePix

KW - Semantic content

KW - Semantic indexing

KW - Visual content

UR - http://www.scopus.com/inward/record.url?scp=8844279974&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=8844279974&partnerID=8YFLogxK

U2 - 10.1117/12.527316

DO - 10.1117/12.527316

M3 - Article

AN - SCOPUS:8844279974

VL - 5292

SP - 414

EP - 425

JO - Scanning Electron Microscopy

JF - Scanning Electron Microscopy

SN - 0586-5581

ER -