Highly Consistent Sequential Segmentation

Michael Donoser, Martin Urschler, Hayko Riemenschneider, Horst Bischof

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

This paper deals with segmentation of image sequences in an unsupervised manner with the goal of getting highly consistent segmentation results from frame-to-frame. We first introduce a segmentation method that uses results of the previous frame as initialization and significantly improves consistency in comparison to a single frame based approach. We also find correspondences between the segmented regions from one frame to the next to further increase consistency. This matching step is based on a modified version of an efficient partial shape matching method which allows identification of similar parts of regions despite topology changes like merges and splits. We use the identified matched parts to define a partial matching cost which is then used as input to pairwise graph matching. Experiments demonstrate that we can achieve highly consistent segmentations for diverse image sequences, even allowing to track manually initialized moving and static objects.
LanguageEnglish
Title of host publicationImage Analysis
Subtitle of host publication17th Scandinavian Conference, SCIA 2011, Ystad, Sweden, May 2011. Proceedings
EditorsAnders Heyden, Fredrik Kahl
PublisherSpringer International Publishing AG
Pages48-58
Volume6688
ISBN (Electronic)978-3-642-21227-7
ISBN (Print)978-3-642-21226-0
DOIs
StatusPublished - 2011
EventScandinavian Conference on Image Analysis - Ystad Saltsjöbad, Sweden
Duration: 23 May 201127 May 2011

Publication series

NameLecture Notes in Computer Science

Conference

ConferenceScandinavian Conference on Image Analysis
CountrySweden
CityYstad Saltsjöbad
Period23/05/1127/05/11

Fingerprint

Topology
Costs
Experiments

Fields of Expertise

  • Information, Communication & Computing

Cite this

Donoser, M., Urschler, M., Riemenschneider, H., & Bischof, H. (2011). Highly Consistent Sequential Segmentation. In A. Heyden, & F. Kahl (Eds.), Image Analysis: 17th Scandinavian Conference, SCIA 2011, Ystad, Sweden, May 2011. Proceedings (Vol. 6688, pp. 48-58). (Lecture Notes in Computer Science). Springer International Publishing AG . DOI: 10.1007/978-3-642-21227-7_5

Highly Consistent Sequential Segmentation. / Donoser, Michael; Urschler, Martin; Riemenschneider, Hayko; Bischof, Horst.

Image Analysis: 17th Scandinavian Conference, SCIA 2011, Ystad, Sweden, May 2011. Proceedings. ed. / Anders Heyden; Fredrik Kahl. Vol. 6688 Springer International Publishing AG , 2011. p. 48-58 (Lecture Notes in Computer Science).

Research output: Chapter in Book/Report/Conference proceedingChapter

Donoser, M, Urschler, M, Riemenschneider, H & Bischof, H 2011, Highly Consistent Sequential Segmentation. in A Heyden & F Kahl (eds), Image Analysis: 17th Scandinavian Conference, SCIA 2011, Ystad, Sweden, May 2011. Proceedings. vol. 6688, Lecture Notes in Computer Science, Springer International Publishing AG , pp. 48-58, Scandinavian Conference on Image Analysis, Ystad Saltsjöbad, Sweden, 23/05/11. DOI: 10.1007/978-3-642-21227-7_5
Donoser M, Urschler M, Riemenschneider H, Bischof H. Highly Consistent Sequential Segmentation. In Heyden A, Kahl F, editors, Image Analysis: 17th Scandinavian Conference, SCIA 2011, Ystad, Sweden, May 2011. Proceedings. Vol. 6688. Springer International Publishing AG . 2011. p. 48-58. (Lecture Notes in Computer Science). Available from, DOI: 10.1007/978-3-642-21227-7_5
Donoser, Michael ; Urschler, Martin ; Riemenschneider, Hayko ; Bischof, Horst. / Highly Consistent Sequential Segmentation. Image Analysis: 17th Scandinavian Conference, SCIA 2011, Ystad, Sweden, May 2011. Proceedings. editor / Anders Heyden ; Fredrik Kahl. Vol. 6688 Springer International Publishing AG , 2011. pp. 48-58 (Lecture Notes in Computer Science).
@inbook{8093d912b896453d8071fb9864006590,
title = "Highly Consistent Sequential Segmentation",
abstract = "This paper deals with segmentation of image sequences in an unsupervised manner with the goal of getting highly consistent segmentation results from frame-to-frame. We first introduce a segmentation method that uses results of the previous frame as initialization and significantly improves consistency in comparison to a single frame based approach. We also find correspondences between the segmented regions from one frame to the next to further increase consistency. This matching step is based on a modified version of an efficient partial shape matching method which allows identification of similar parts of regions despite topology changes like merges and splits. We use the identified matched parts to define a partial matching cost which is then used as input to pairwise graph matching. Experiments demonstrate that we can achieve highly consistent segmentations for diverse image sequences, even allowing to track manually initialized moving and static objects.",
author = "Michael Donoser and Martin Urschler and Hayko Riemenschneider and Horst Bischof",
year = "2011",
doi = "10.1007/978-3-642-21227-7_5",
language = "English",
isbn = "978-3-642-21226-0",
volume = "6688",
series = "Lecture Notes in Computer Science",
publisher = "Springer International Publishing AG",
pages = "48--58",
editor = "Anders Heyden and Fredrik Kahl",
booktitle = "Image Analysis",
address = "Switzerland",

}

TY - CHAP

T1 - Highly Consistent Sequential Segmentation

AU - Donoser,Michael

AU - Urschler,Martin

AU - Riemenschneider,Hayko

AU - Bischof,Horst

PY - 2011

Y1 - 2011

N2 - This paper deals with segmentation of image sequences in an unsupervised manner with the goal of getting highly consistent segmentation results from frame-to-frame. We first introduce a segmentation method that uses results of the previous frame as initialization and significantly improves consistency in comparison to a single frame based approach. We also find correspondences between the segmented regions from one frame to the next to further increase consistency. This matching step is based on a modified version of an efficient partial shape matching method which allows identification of similar parts of regions despite topology changes like merges and splits. We use the identified matched parts to define a partial matching cost which is then used as input to pairwise graph matching. Experiments demonstrate that we can achieve highly consistent segmentations for diverse image sequences, even allowing to track manually initialized moving and static objects.

AB - This paper deals with segmentation of image sequences in an unsupervised manner with the goal of getting highly consistent segmentation results from frame-to-frame. We first introduce a segmentation method that uses results of the previous frame as initialization and significantly improves consistency in comparison to a single frame based approach. We also find correspondences between the segmented regions from one frame to the next to further increase consistency. This matching step is based on a modified version of an efficient partial shape matching method which allows identification of similar parts of regions despite topology changes like merges and splits. We use the identified matched parts to define a partial matching cost which is then used as input to pairwise graph matching. Experiments demonstrate that we can achieve highly consistent segmentations for diverse image sequences, even allowing to track manually initialized moving and static objects.

U2 - 10.1007/978-3-642-21227-7_5

DO - 10.1007/978-3-642-21227-7_5

M3 - Chapter

SN - 978-3-642-21226-0

VL - 6688

T3 - Lecture Notes in Computer Science

SP - 48

EP - 58

BT - Image Analysis

PB - Springer International Publishing AG

ER -