Highly Consistent Sequential Segmentation

Michael Donoser, Martin Urschler, Hayko Riemenschneider, Horst Bischof

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review


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.
Original 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
ISBN (Electronic)978-3-642-21227-7
ISBN (Print)978-3-642-21226-0
Publication statusPublished - 2011
EventScandinavian Conference on Image Analysis - Ystad Saltsjöbad, Sweden
Duration: 23 May 201127 May 2011

Publication series

NameLecture Notes in Computer Science


ConferenceScandinavian Conference on Image Analysis
CityYstad Saltsjöbad

Fields of Expertise

  • Information, Communication & Computing


Dive into the research topics of 'Highly Consistent Sequential Segmentation'. Together they form a unique fingerprint.

Cite this