Skip to main content
Log in

Bottom-up segmentation of image sequences for coding

Segmentation de bas en haut de sÉquences d’images en vue du codage

  • Published:
Annales Des Télécommunications Aims and scope Submit manuscript

Abstract

A bottom-up segmentation method is presented as a first step of object based coding. An initial partition is created based on connected filters. The iterative fusion algorithm merges the most similar regions until the target is reached. As the regions become bigger, more complex criterion, like similarity of texture or of motion are used. The algorithm offers a good balance between time stability and the ability to cope with the apparition of new regions.

Résumé

Une méthode de segmentation de bas en haut est présentée, destinée à servir de première étape d’un codeur orienté objet. Grace à des filtres connexes, une partition initiale est créée. Un algorithme de fusions itératives réunit les régions les plus semblables tant qu’un critère d’arrêt n’est pas vérifié. A mesure que les régions atteignent des tailles plus importantes, des critères plus complexes, tels que la similarité des textures ou du mouvement sont utilisés. L’algorithme offre un bon compromis entre stabilité temporelle et capacité d’introduire de nouvelles régions dans la segmentation.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

REFERENCES

  1. Crespo (J.). Morphological connected filters and intra-region smoothing for image segmentation.PhD Thesis, School of Electrical Engineering, Georgia Institute of Technology (1993).

  2. Gilge (M.), Engelhardt (T.), Mehlan (R.). Coding of arbitrarily shaped image segments based on a generalized orthogonal transform.Image Communication (1989), 1, n° 2.

  3. LEP. First results of motion prediction based on a differential estimation method.In SIM(94) 35 COST 211, Tampere (June 1994).

    Google Scholar 

  4. Marcotegui (B.), Marqués (F.), Morros (R.), Pardà;s (M.), Salembier (P.). Segmentation of video sequences and rate control.Ann. Télécommunic. (1997), 52, n° 7-8, pp. 380–388.

    Google Scholar 

  5. Meyer (F.). Algorithmes à; base de files d’attente hiérarchique.Technical Report NT-46/90/MM, Ecole des Mines de Paris, Centre de Morphologie Mathématique (Sep. 1990).

  6. Meyer (F.), Marqués (F.). Segmentation in object based coding : general requirements.Ann. Télécommunic}. (1998) (à; paraître).

  7. Meyer (F.), Oliveras (A.), Salembier (P.), Vachier (C). Morphological tools for segmentation: connected filters and watersheds.Ann. Télécommunic. (1997), 52, n° 7-8, pp. 367–379.

    Google Scholar 

  8. Monga (O.). Segmentation d’images par croissance hiérarchique de régions.PhD Thesis, Université Paris Sud, Centre d’Orsay (1988).

    Google Scholar 

  9. Pardà;s (M.), Salembier (P.). Time-recursive segmentation of image sequences.In EUS1PCO-94, Edinburgh (Sep. 1994).

    Google Scholar 

  10. Pardà;s (M.), Salembier (P.). Segmentation of video sequences for partition tree generation.Ann. Télécommunic. (1997), 52, n° 7-8, pp. 389–398.

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding authors

Correspondence to Beatriz Marcotegui or Fernand Meyer.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Marcotegui, B., Meyer, F. Bottom-up segmentation of image sequences for coding. Ann. Télécommun. 52, 397–407 (1997). https://doi.org/10.1007/BF02998459

Download citation

  • Received:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF02998459

Key words

Mots clés

Navigation