From region based image representation to object discovery and recognition

Narendra Ahuja, Sinisa Todorovic

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper presents an overview of the work we have done over the last several years on object recognition in images from region-based image representation. The overview focuses on the following related problems: (1) discovery of a single 2D object category frequently occurring in a given image set; (2) learning a model of the discovered category in terms of its photometric, geometric, and structural properties; and (3) detection and segmentation of objects from the category in new images. Images in the given set are segmented, and then each image is represented by a region graph that captures hierarchy and neighbor relations among image regions. The region graphs are matched to extract the maximally matching subgraphs, which are interpreted as instances of the discovered category. A graph-union of the matching subgraphs is taken as a model of the category. Matching the category model to the region graph of a new image yields joint object detection and segmentation. The paper argues that using a hierarchy of image regions and their neighbor relations offers a number of advantages in solving (1)-(3), over the more commonly used point and edge features. Experimental results, also reviewed in this paper, support the above claims. Details of our methods as well of comparisons with other methods are omitted here, and can be found in the indicated references.

Original languageEnglish (US)
Title of host publicationStructural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings
Pages1-19
Number of pages19
DOIs
StatePublished - Oct 29 2010
Event7th Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition, SSPR and SPR 2010 - Cesme, Izmir, Turkey
Duration: Aug 18 2010Aug 20 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume6218 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other7th Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition, SSPR and SPR 2010
CountryTurkey
CityCesme, Izmir
Period8/18/108/20/10

Fingerprint

Image Representation
Object recognition
Structural properties
Graph in graph theory
Subgraph
Segmentation
Model Category
Object
Object Detection
Object Recognition
Structural Properties
Union
Experimental Results

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Ahuja, N., & Todorovic, S. (2010). From region based image representation to object discovery and recognition. In Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings (pp. 1-19). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6218 LNCS). https://doi.org/10.1007/978-3-642-14980-1_1

From region based image representation to object discovery and recognition. / Ahuja, Narendra; Todorovic, Sinisa.

Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings. 2010. p. 1-19 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6218 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ahuja, N & Todorovic, S 2010, From region based image representation to object discovery and recognition. in Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 6218 LNCS, pp. 1-19, 7th Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition, SSPR and SPR 2010, Cesme, Izmir, Turkey, 8/18/10. https://doi.org/10.1007/978-3-642-14980-1_1
Ahuja N, Todorovic S. From region based image representation to object discovery and recognition. In Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings. 2010. p. 1-19. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-14980-1_1
Ahuja, Narendra ; Todorovic, Sinisa. / From region based image representation to object discovery and recognition. Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings. 2010. pp. 1-19 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{83bc96272b2549feaf384ae36b4fef36,
title = "From region based image representation to object discovery and recognition",
abstract = "This paper presents an overview of the work we have done over the last several years on object recognition in images from region-based image representation. The overview focuses on the following related problems: (1) discovery of a single 2D object category frequently occurring in a given image set; (2) learning a model of the discovered category in terms of its photometric, geometric, and structural properties; and (3) detection and segmentation of objects from the category in new images. Images in the given set are segmented, and then each image is represented by a region graph that captures hierarchy and neighbor relations among image regions. The region graphs are matched to extract the maximally matching subgraphs, which are interpreted as instances of the discovered category. A graph-union of the matching subgraphs is taken as a model of the category. Matching the category model to the region graph of a new image yields joint object detection and segmentation. The paper argues that using a hierarchy of image regions and their neighbor relations offers a number of advantages in solving (1)-(3), over the more commonly used point and edge features. Experimental results, also reviewed in this paper, support the above claims. Details of our methods as well of comparisons with other methods are omitted here, and can be found in the indicated references.",
author = "Narendra Ahuja and Sinisa Todorovic",
year = "2010",
month = "10",
day = "29",
doi = "10.1007/978-3-642-14980-1_1",
language = "English (US)",
isbn = "3642149790",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "1--19",
booktitle = "Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings",

}

TY - GEN

T1 - From region based image representation to object discovery and recognition

AU - Ahuja, Narendra

AU - Todorovic, Sinisa

PY - 2010/10/29

Y1 - 2010/10/29

N2 - This paper presents an overview of the work we have done over the last several years on object recognition in images from region-based image representation. The overview focuses on the following related problems: (1) discovery of a single 2D object category frequently occurring in a given image set; (2) learning a model of the discovered category in terms of its photometric, geometric, and structural properties; and (3) detection and segmentation of objects from the category in new images. Images in the given set are segmented, and then each image is represented by a region graph that captures hierarchy and neighbor relations among image regions. The region graphs are matched to extract the maximally matching subgraphs, which are interpreted as instances of the discovered category. A graph-union of the matching subgraphs is taken as a model of the category. Matching the category model to the region graph of a new image yields joint object detection and segmentation. The paper argues that using a hierarchy of image regions and their neighbor relations offers a number of advantages in solving (1)-(3), over the more commonly used point and edge features. Experimental results, also reviewed in this paper, support the above claims. Details of our methods as well of comparisons with other methods are omitted here, and can be found in the indicated references.

AB - This paper presents an overview of the work we have done over the last several years on object recognition in images from region-based image representation. The overview focuses on the following related problems: (1) discovery of a single 2D object category frequently occurring in a given image set; (2) learning a model of the discovered category in terms of its photometric, geometric, and structural properties; and (3) detection and segmentation of objects from the category in new images. Images in the given set are segmented, and then each image is represented by a region graph that captures hierarchy and neighbor relations among image regions. The region graphs are matched to extract the maximally matching subgraphs, which are interpreted as instances of the discovered category. A graph-union of the matching subgraphs is taken as a model of the category. Matching the category model to the region graph of a new image yields joint object detection and segmentation. The paper argues that using a hierarchy of image regions and their neighbor relations offers a number of advantages in solving (1)-(3), over the more commonly used point and edge features. Experimental results, also reviewed in this paper, support the above claims. Details of our methods as well of comparisons with other methods are omitted here, and can be found in the indicated references.

UR - http://www.scopus.com/inward/record.url?scp=77958472064&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77958472064&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-14980-1_1

DO - 10.1007/978-3-642-14980-1_1

M3 - Conference contribution

AN - SCOPUS:77958472064

SN - 3642149790

SN - 9783642149795

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 1

EP - 19

BT - Structural, Syntactic, and Statistical Pattern Recognition - Joint IAPR International Workshop, SSPR and SPR 2010, Proceedings

ER -